Apple's iOS 26 Review, Meta to Pay $15B for Scale AI Stake | Andrew Huberman, Scott Belsky, Sai Senthilkumar, Karri Saarinen, Roy Bahat, Alex Israel

You're watching TVPN. Today is Tuesday, June 10, 2025. We are live from the TVPN Ultra Dome, the temple of technology, the fortress of finance, the capital of capital. It's a great day. We're doing WWC reactions. I will run you through a little bit of the show. We want to talk about the meta scale AI deal. 49% for just $14 billion. It's a steal. You're getting Alex, you're getting the Scale AI team. They're giving it away. We're going to talk about the Glean fund raise competitor OpenAI. They're cooking in enterprise AI. Some of the reaction there from what's going on in deep research and some of the advancements that OpenAI has been rolling out there. Very interesting for Glean's business. Glean still ripping. So they're up, raised a big round to today. They also got the cnbc disruptor list. 50 top startups. We've had a ton of them on the show. We're going to go through do a little underrated, overrated maybe see who should be higher. John, personally I love them all. I think they should all be tied for first place. But CNBC ranked them 1 to 50.

Hopefully it was purely pay to play and CNBC just kind of auctioned off the spots.

That's.

That would be the most fair thing, you know.

Agree, agree.

But.

And then we got a bunch of great guests. We got CY from Redpoint coming on. We have Kari announcing a massive. What is a series C unicorn round at linear one of our.

It must be their Series C. We.

Got Belsky from a 24 coming on. Roy Behat from Bloomberg Beta coming on, announcing a new $75 million fund. And then this guy, and then Andrew. Andrew.

This guy, Andrew Huberman.

Andrew.

He may have heard of him. Fellow podcaster.

Yeah. So we're excited for the show today, but let's kick it off with the timeline's reaction to wwdc. Liquid glass. Whatever says Juan. The most important thing in iOS 26 text is finally left aligned. You know, I saw people posting about this. You said this about like the center alignment, left alignment thing. I hadn't noticed it. It looks a lot better left aligned. I don't know why. I didn't know what I was missing.

We'll take it.

Huge improvement.

We'll take it.

Yeah. So now text in iOS notifications, left aligned. Huge news. Also all the different version numbers are now they're doing car models. So it's iOS 26. They jump forward a bunch. I Believe we were on like 18 before, so they just skipped a bunch.

I can't wait to be in the thousands.

We're in 20, 25 in the thousands and it's the iOS 26 across everything. So unification there. It's happened in Android a while ago, but now it's come to iOS.

Love it.

Again, a lot of people with negative reactions. We're going to take you through some of the positive reactions. Steve Jobs would have cornered you in a dark alley and beat you with a metal pipe. He suggested something like this. That's what Kitsy says. This was yesterday's reaction. Today some of the designers are coming out and saying, hey, let them cook. There's a silver lining here.

There's certain examples that are really, really gnarly.

So there's that one super viral one that's like really hard to see. And then people put it next to the Oppenheimer with the blast. I think that's a fake photo. I think they adjusted the brightness on that. I don't think they. No, not a deep fake. I don't think there's any AI involved. I just think that someone went in there and was like, let me turn off the brightness for effect. And I think that that's fake. And I think that also, again, this is, this is a beta. This is something that they're going to be iterating on before it goes out to the masses.

It's honestly a bit weird for a trillion dollar company to launch a beta. They should just say, no, this is an alpha.

Yes, but, but, but it makes sense in terms of like, like if you are a developer and there's a, and there's a significant change to the user interface, you want to bake that into your app so that you don't look out of place on day one when everyone updates.

Yeah.

So you gotta get the beta into the hands of the developers so that the Airbnb app and the Public.com app and the Ramp.com app. Save time and money. Time is money. Save Both easy to use. Corporate cards, bill payment, accounting and a whole lot more all in one place. Go to ramp.com and travel.

John's favorite.

And travel. The travel.

Favorite thing.

Speaking of which, we got to book our, we got to book our hotels for tonight because we're going to San Francisco for YC demo day tomorrow.

That's right.

It's going to be a massive stream.

With Ramp Travel, you, you can book last minute and it's still a breeze.

It's really, really easy. But yes, if you're designing an app, you want it to feel somewhat seamless with the rest of the ui. And if there's UI changes, you got to test drive those beforehand. You got to make sure that those are right. So the big question. Should we have our intern Tyler Cosgrove over there install iOS 26 beta? You know, it's potentially. He's on the green screen. Something happened. He's on the green screen. Should we make Tyler install? Oh, there. There we go. That's where he is. Really? Like, I'm glad that you did the false background for a minute and make it look like a green screen instead of showing the real.

Yeah, we wanted him to be more approachable.

Yes. We just put a green screen behind him as a fake background.

Yeah, but, but, but anyways, back to you. Is that Mykonos? Where are you right now?

Oh, that's just my. My pool in my backyard.

Yeah.

Okay, cool.

This is where you take.

Sometimes fellow. Some of my friends will go on podcasts. They'll just use my backyard.

Yeah. Yeah. O. That's what it is.

Anyways, I would love to watch you risk it all and install the update. So why don't you.

Are you willing to do it?

I'll do it. Yeah.

Okay, do it.

He's gonna do it for the team. He's taking one for the team.

IOS 26 beta. It could go terribly wrong because the bugs. You know, the thing with the beta is that you can't go back. You can't roll back to the non beta.

Yeah.

And so you could be living with minor grading bugs for, like, months until they fix it now. They fix it pretty quickly. I've done it, like, once or twice. Gone on the beta because I'm like, oh, I'm so excited. I want to try the new thing.

This is basically intern hazing.

It is hazing in the modern era. The WWC version is set up.

All right, Tyler, well, you are brave.

Okay.

Good luck with that and why don't you get started?

Get started and then give us the review. We'll be checking in with you to see how the iOS 26 installation process goes and. And any problems you run into there. We also have some other.

I can't read anything on my phone anymore.

We also have some big show updates. Not only do you have intern cam now, we also have the breaking news camera. Let's go to it. Now.

This is crazy, too, because I can.

See you see this. Okay. This is the breaking news camera. When we have breaking news, we cut to the printer. Cut, cut.

And and it prints out.

You can read the breaking news. Okay, Jordan, you want to read that to us? Hold it up.

We have some breaking news. If we can switch cameras.

What is it?

So, breaking news. Snap to launch smaller lighter augmented reality specs smart glasses in 2026. I'm excited for this from CNBC. Thank you to CNBC. Snap on Tuesday announced its plans to release a sixth generation of its augmented reality glasses in 2026. As competition in the smart glasses market continues to heat up. The the maker of Snapchat said that its next generation glasses will be called Specs. Breaking with the company's Spectacles branding that it used for previous Drop the ACLs. Drop the ACLs just use specs. It's cleaner.

Well, they drop the T too. So it's technically Drop the tackles.

Drop the tackles. Just Specs. Larger tech rivals like Meta, Apple and Alphabet are investing heavily into cutting edge head worn devices.

Yep.

This computer will be on your face.

You're most recent. The most recent edition of Spectacles were released in September 2024 to developers only. That edition of the glasses was available under a leasing model that required users to commit to paying 99amonth for a full year if you wanted to develop for them. So that's $1,200 guaranteed. And then if you want to keep it more, it's probably continues to be $100. So it'll be interesting. It'll run Snap OS operating system which they develop. Snap said the developers will be able to incorporate Google's Gemini AI models into programs that they develop for smart glasses. So that's interesting. Some sort of. Yeah, exactly. So you could see someone where it's like okay, you're going to get the glasses, you're going to build some.

Google says, hey, oh, that's where you're going with it.

Okay.

I don't know. I, I think very unlikely but, but still, you know.

Well, Snaps market cap, exactly what Meta just paid for a 50% stake in scale AI. 14 billion. Anything's possible.

Spicy.

So SNAP's at 14 billion. The, the stock hasn't really moved over the past five days on this news. It's clearly Snap, you know, staying in the game with the WWC news, getting something, getting a story out there that hey, we're still in the game, we're moving and we're doing cool stuff and I'm excited for this. I think we need more, more innovation here, more people testing different stuff. And it's clear that with a new format, with a new paradigm of like what the product is there's a whole bunch of interesting directions you can go whether there's an augmented reality display, whether it's voice only, whether it's glasses.

It's crazy that the glasses.

A whole bunch of different things.

Social media goats, both of them are just not. The idea of not being a part of the next platform shift is just an idea. They don't. Is a world they don't want to live in. Right. They are. You know, both Zuck and Evan are gonna be, you know, competing.

Yeah. And so fun. Could be interesting from a developer angle. Obviously, you know, nowhere near the budget of Reality Labs for what Zuck is working on with Meta. But potentially constraints.

It can be good, John.

I agree. I agree.

They can be good.

I'm optimistic. I certainly want to try it anyway. Let's go back to Liquid Glass. Ben Thompson gave a whole posted a massive article on strutechery called Apple's Retreat. Apple Retreats. Something like that. Should give it a read. You should subscribe to Strikery obviously. Apple's retreat to its core competency. The headline feature of WWDC this year. This is from the latest structecary article was Liquid Glass, a new unified design language that stretches across its operating systems. I will reserve judgment on liquids, glass aesthetics and usability. Gruber likes it, but I am journalists on the horizon. But I am not one to install developer betas on my devices.

But Tyler will.

But Tyler will. That's why we have Tyler. Because I agree with Ben Thompson. It's extremely risky, but he's willing to do it. Taking one for the team over there.

I mean the whole. The whole TVPN army. He's. He's really saying, I will, I will sacrifice.

I will fall on the sword.

I will fall on the sword.

It's okay if it goes really south, we'll get it. We'll get a new phone. And it's always, always a possibility. But good luck. Hopefully you make it through. Tyler will be checking on you throughout the show. So first let's read Gruber's. Gruber's overview of his reaction to Liquid Glass. Because we saw Timeline, hated it. They said Steve Jobs would hate it. Let's go to John Cooper. Darren Fowler.

I think part of it is it was popular to hate on it.

Totally.

And everyone loves.

Anytime someone's on a downswing, they love just piling on downwards. And when everyone's on an upswing, they love just piling on and upwards. It's not.

I think, I think it's also normal. If a widely Used consumer product does any type of meaningful upgrade. I mean, this default is that there.

Were protests that it's bad or Facebook back in the day launching the feed. Yeah, like I want to go back to just randomly looking at your profile. I don't want to see random all everything. Every change is always met with some sort of pushback. And then years later, people wind up loving it. I feel like this is something that could grow on me. I'm reserving until we've fully tested it on our intern. If he gives it the thumbs up, I'm gonna install it. Let's go to John Gruel.

I mean, it's kind of fun. It's daring and brave and bold to do something like that. And just because Tyler's doing it doesn't mean you shouldn't.

This type of bravery that Tyler's exhibiting today, I think it might be one day an exhibit in the Computer History Museum. Totally bravery he's exhibiting today will be in the history.

They might even name the next operating system. Cosgrove. Cosgrove.cosgrove.Mac os27mac os cosgrove after Tahoe. It's got an amazing ring to it. It's got a nice ring and it just loads with this image. Like, that's the default wallpaper for the.

Daring macroeconologist who installed the beta on day two.

You did. You know what everybody said? You're crazy.

You're crazy.

It's impossible.

You're crazy.

You should never do that. But he did it. He took the leap.

He took the leap and he did tell the tale.

So first, this is going back to Ben Thompson's piece. First, handcrafted UI overhauls are the polar opposite of a probabilistic world of generative AI. One is about deep consideration and iteration resulting in a finished product. The other is about in the moment token prediction, resulting in an output that is ephemeral and disposable. Both are important and creative. But the downsides of that creativity, unfamiliarity and edge cases versus hallucination and false confidence are themselves diametrically opposed. Apple's historical strengths have always been rooting and designing for finality. In my first year of shirtechery, I did a SWOT analysis. Let's give it up for swot. I love SWOT analysis of the big tech companies and said about Apple and I think we've talked about this a bunch on the show of like generative AI is inherently imperfect. Apple is about pixel perfection. So Apple's product development process is wonderful for developing finished products, but that same Process doesn't work nearly as well when it comes to building cloud services. Cloud services are never done. They are best developed by starting with an MVP and then iterating based on usage. This is precisely opposite of what it takes to design a piece of hardware. And it's a big reason why Apple struggles so much with cloud services and why other services companies struggle with products. The canonical example of this of course was when Mobile Me was the Mobile Me launch, which was delivered fully formed and which when faced with real world usage, crash and burned. Apple's latest offerings are better, but still suffer from too much internal development time per release. This is a hard problem to fix because it touches the core of what makes Apple Apple.

I think it matters. I think it matters whether or not liquid glass is good because it will be a testament about the state of Apple strengths. The point of this article however, is that WWC was a retreat to those strengths, which is good. This is a bull case. They're not getting over their skis and, and they will eventually. I mean if they're to win, probably defer to the other companies and play to their strengths and let the other companies win in the places where they are.

Tim just calling up Sam and saying, sam, I'm going to need 50 billion.

Siri, it's yours.

50 billion, 50 billion. You know, if you and Masa can pull together the capital, we can make it happen.

I mean, I keep thinking about Siri and I keep running through like what would it have taken to, to just, to just replace Siri with a frontier level, you know, simple response model, not even the Reasoning stuff, just GPT4O level, you know, you ask it a question and it, and it, and it reads you like a basic response and not lose the other Siri functionality. And I think that it need, they need to kind of flip the paradigm because right now Siri is only, it's basically all tool use, it's basically a router. So it takes in, you know, you ask Siri what what's the weather going to be like tomorrow? And it knows that you just. And all it's doing is picking out keywords weather and it goes to the weather API internally and then tomorrow and it feeds that in and then it delivers that back. Like ChatGPT kind of flips that around by having the default interaction be much more generalizable and it can return kind of any result, it can return a paragraph, it can tell you about a historical event or just talk to you. And then if it needs to use a tool like web search to get you the weather for tomorrow. Then it goes and calls that. And so you need to kind of flip that around. I was wondering about like, like, why hasn't, like, what, what if you, if you rolled back the clock and you were like, okay, Apple just wants to be in the game, what would it take to launch a V2 of Siri? And I was like, well, okay, if they try to partner with someone under the hood without, like, they don't even have the option of just saying, we are going to partner with Anthropic under the hood and it's going to be, and it's going to be basically white labeled and no one will ever find out because there are just too many leakers. Like it will find out immediately. And so Mark Gurman will have the story immediately and everyone will know. So they won't be able to say like this is magically ours when really it's someone else's, even if they're paying them to be. And there's tons of NDAs like they.

Do with cloud services. Right. I don't actually know who under.

I think they, I think they do a lot of that. And then I think a lot of it comes down to the privacy angle is that they want to make certain claims about privacy and ESG and NET zero and stuff. And so if they were to just white label another company that didn't have the same reputation in privacy or, or environmentalism, like people would just say, wait, wait, wait. But like Anthropic isn't, isn't known for their privacy rules as much as possible. So maybe they're training on my Siri results now and that's a whole, that's a whole thing. And they have to, you know, build up Anthropic's brand to have the brand of Apple, which is just, I'm sure Anthropic's doing great, but they're not known like Apple is in terms of how seriously they take it. Yeah. And then in terms of, in terms of acquisition, obviously there's all the antitrust stuff. We'll get into this with a scale AI deal. But every major hyperscaler except for Apple now has basically done one of these like interesting aqua hires of a foundation model lab to kind of juice up their AI efforts. And, and, and it's kind of been a mixed bag. Maybe we see one from Apple. They certainly have the cash, but it would be a very, very different, it would be a departure from their current strategy. Anyway, let's go to Jon Gruber over at Daring Fireball with his review of Liquid Glass, he says, I've got iOS 26 installed on a spare phone already. Oh no. He didn't even use the real phone. Tyler, mistake number one, never download on your main phone. How's it going so far?

Well, okay, so I'm going to download it. It's 19 gigabytes. Okay, so my phone is not, you know, I only have 64 gigabytes on my phone. So I'm going to, I gotta do.

A, you know, Delete everything.

Yeah, I'm gonna delete everything.

Okay.

Yeah, yeah, yeah. So delete. Just go to your photos app, select all the family photos and then clear the trash and you'll make more memories. Like memory you could always make. You're making more memories all the time.

Yeah, exactly.

Clear out those images.

Good luck.

On a more serious note, if you need, we'll cover your sort of Apple icloud plan if you need to get those images in the cloud. So you can get, if you need.

A backup on your ramp card.

Yeah, put it on the ramp.

So Gruber says, I like the new UI a lot. In addition to just plain looking cool, Apple has tackled a lot of long standing minor irritants. For example, the iOS contextual menu for text selections, the one with cut, copy, paste. For years there, for years now. There have been a lot of other useful commands in there, including share at the very end. But to get to the extra commands you had to tediously swipe, swipe, swipe. Now with one tap you can expand the whole thing into a vertical menu. Elegant. And I agree, I noticed that that was getting like. It was originally it was just like just copy, just, just cut and paste and it was pretty easy. And then they added if you wanted to bold or italics, you had to swipe over and then click on that and then go to text. And there were like seven different things. Eventually, yeah, good, good move, minor update, but important. And this is what this is. This is the polish that Apple's so good at. And then he says there's some stuff in Mac ox mass, Mac OS 26 Tahoe I already don't like. Like putting needless icons next to almost every single menu item. But overall my first impression of Liquid Glass on Mac OS is good too. It's fun and lots of little details are nice, joyful and useful in an.

Old school macOS Tahoe. You'll love Mac OS Cosgrove.

I'll just say that this is going to be a banger. It's going to be the best ever. There's an interesting article in here that Ben Thompson linked to by Sebastian De with about physicality, the new age of UI. He actually posted this back on June 3rd before Apple refreshed this. So there's a lot of. Oh, do we have some breaking news?

I just, I just printed this out because, well, so this was a good post from Signal. He says watching this wwdc, thank you Signal for the, for the content and I'm just happy to be back printing.

It's amazing.

It's, it's. I don't know how we went a few months without it. So he says watching this WWE see felt like watching a dinosaur try to do ballet. Technically and aesthetically competent, but spiritually, spiritually lost. They talked about Apple intelligence like they just discovered the concept throughout the presentation. There was zero ecosystem thinking and no real attempt to rethink a single interaction model. More importantly, there was zero sense of how to be relevant. In a world where interfaces may disappear, apps get abstracted, and AI is the os. Apple still thinks in icons and widgets. While the world is moving towards awareness, intent and delegation, it'll be interesting to see how much of a penalty they will pay for missing this boat.

Another negative post. Now Signal did have a couple other posts that were more positive.

Talking about good stuff. I think this is generally why I'm excited about Johnny and Sam's hardware project is it's an opportunity to rethink the device from the ground up with AI at the heart versus Apple, which I think going back to Ben Thompson's position, is smart to play their own game right now and understand that they do have the option of auctioning the sort of Siri out to the highest bidder at some point.

Yeah, I mean, Alex Heath, who we had on the show yesterday, had a good post who said this is so very clearly designed for an interface that sits on top of the real world, not a phone, something you would need for AR glasses. And so this idea that it's a little awkward now when you see these images of these glass over the background of your phone display, it can look a little too transparent. The contrast ratios are all off. But if you think about if they're really going all in on AR VR and they want to project these screens over the real world long term, setting themselves up to really dominate there and not need to do another refresh. Could be big. Could be big. Anyway, Mark Gurman liked it.

Spent some time with Trump. Yeah, Trump is known for playing 4D chess. Maybe Tim is playing a little 4D chess with Liquid glass. He's like screen is going to be, everything's going to be a screen. We're just going to put it.

They have a patent for a glass iPhone that, that you can see through or something like that or, or something that actually would appear more like glass or appear transparent. And a lot of the transparent, I feel like a lot of the transparent technologies, it's like the closest we're going to get to just true holograms. If you have a variety of devices that are projecting holograms into glass sheets. And so maybe it doesn't start with your phone immediately because the battery needs to be right there, that needs to be solid. But if you think about a display and you're projecting on that and doing kind of the same wave guides that are happening in reflected augmented reality glasses, you could do that on a monitor or something like that. And so, so that could be very cool. Mark Gurman had a more positive take on wwdc. He says, excellent wwdc, cohesive story, deep integration and continuity across the devices, zero false promises, impressive new UI and significant new productivity features on the Mac and iPad. But the lack of any real new AI features, despite that being my expectation, is startling. And so people were expecting, you know, maybe there will be a new Siri, maybe they will move the ball down the field in AI and they did not. Ben Hilak, of course, taking credit for it as always. You know him from taking credit for the Jaguar rebrand now he says. So excited to see my, my, to finally see my work from Apple ship. I was the head of design for the WWDC app notifications. There's still a work in progress. We've only had the last year to work on them. Stay tuned.

So was this, was this, is this fake or is this real?

I think this is brightness adjusted. I think that they, I think that whoever shared this image turned up the brightness or brought up the, brought up the lows, basically decreased the contrast in the, in the Photos app. Like to make it look worse than it actually would. But I mean even it doesn't even matter because even if this is the way it looks, this isn't the way it will look when it will ship to people because it's obvious and Apple doesn't make these obvious mistakes like this, especially when it comes to design. They might do something that is like rough in the short term, but better in the long term. Like what I think is happening with the Photos app, where in the Photos app they are understanding that in the future you won't want to scroll at all. You will just search and search will be so powerful because every photo and video will be tagged so thoroughly from AI and it will understand the context so much that you will just ask it to pull up a photo and it will do it immediately. The problem is is that right now people a aren't they haven't switched that UX paradigm to actually go search first.

Yeah.

And second and second just the tagging is not that good. Like I had a video clip of Arnold Schwarzenegger from a movie on my phone. I wanted to pull it up and so I searched like Arnold Arnold video and it couldn't find it because the clip that I had started with a black frame and clearly it's only indexing like the first frame. And so if you have like a fade in to a picture of your dog and it won't get text, something like that, I don't know if that's exactly what's happening.

I use, I use the text search in photos quite a bit.

Yep.

I find it like searching Coogan. If there's some argument that has.

Oh totally. The text search is fantastic. And even, even things like objects like I. The picture of a vending machine. You type in vending machine. It'll do a pretty good job sometimes. It'll miss it though. But these like generative AI things, they get to 99.9% accuracy pretty quickly. Apple's just a little bit behind on them, I think. Anyway, if you're trying to upgrade your design chops, head over to figma.com think bigger, build faster. Figma helps design and development teams build great products together.

We'll go to Willem on figma. Now we'll go to.

We'll go to Willem. He gives his review of Apple's Liquid Glass. Clearly hints at a clear interface, AR glasses and unifies OS beyond devices like iOS 7. I would be shocked if this version of Liquid Glass ships to everyone later this year. Too many accessibility features for a company that pioneered it. Unfortunately the this means it's going to be muddy slash blurry. UI design used to be about function. This is decoration. Late Empire feeling hate. The trend of floating menus that seems to be happening everywhere. It puts useless noise on the edges of the screen. Ultimately little of this matters when you can now just talk to the computer. Everything is computer anyway.

Interesting.

There's an interesting deep dive from the Sebastian de Witte talking about some of the history here. Speaking of iOS 7, he said on June 10th Apple showed off what this was in the spring of 2013. June 10th Apple showed off what would be the greatest paradigm shift in user Interface Design ever? IOS 7. I remember exactly where I was and how I felt. It was a shock. I love a designer who's just so in the weeds. They remember where they were when iOS7 dropped. If there is indeed a big redesign happening this year, it'll be consequential and impactful in many ways that will dwarf the iOS 7 overhaul for a multitude of of reasons. The redesign is rumored to be comprehensive, a restyling of iOS, macOS, iPadOS, TVOs, WatchOS and Vision OS. In the intervening years, between iOS7's announcement and today, iPhones have gone from simply a popular device to the single most important object in people's lives. So true. The design of iOS affected and inspired most things around, from web to graphic design to any other computer interface. That's why I figured I'd take this moment of obscurity, this precious moment in time where its changes are still shrouded in fog, to savor something wholesome, naivete. Great writer of where things are going, so I can let my imagination run wild. What would I do if I were Apple's design team? What changes would I like to see and what do I think is likely? And then he breaks it down. The shaded age. He goes back to the skeuomorphic era. IOS started out as.

Can we pull up this?

IOS started out as the iPhone OS, an entirely new operating system that had very similar styling and to the design language of the Mac OS Tiger dashboard feature. Yeah, I remember this. You can just put the widgets anywhere. It was very messy. I'll just stick.

This was so fun as a kid.

Yeah, yeah.

Messing around with all the different widgets.

All the widgets. And then you would.

Discovering. Discovering widget. I couldn't. I legitimately felt like a kid in a candy store with widgets. Yeah, it was. It was a beautiful time.

So he says the icon layout on iPhone OS 1, it wasn't even called iOS then. It was just iPhone OS was a clear skeuomorph. Skeuomorphism was the. Was the buzzword at the time. You might have heard that word being thrown around. It might surprise you that it doesn't mean it had lots of visual effects like gradients, gloss and shadows. It actually means that to make it easier for users to transition from something they were used to, in this case, phones typically being slabs with a grid of buttons on them to what they had become. Phones were all screen, so they could show any kind of button or interface imaginable. And yes, there was A whole lot of visual effects in the user interfaces from iPhone OS 1 to iOS 6. In this age, we saw everything from detailed gradients and shadows and simple interface elements to realistically rendered reel to reel tape decks and microphones for audio apps. I remember this, the reel to reel tape deck in the voice recorder. Pretty, pretty remarkable. And like, the directions would actually be like a, like a physical sound sign that you would see on a highway. It was, it was such a cool design language. I mean, it feels so not Apple now that they've spent 10 or 15 years post skeuomorphism, but it was definitely unique at the time. Having actually worked on some of the more fun manifestations of it during my time working at Apple, I can tell you from experience that the work we did in this era was heavily grounded in creating familiarity through thoughtful, extensive visual effects. We spent a lot of time in Photoshop drawing realistically shaded buttons, virtual wood, leather and more. More materials that became known as skeuomorphic design, which I find a bit of a misnomer. But the general idea stands. Of course, the metal of the microphone was not in fact metal. It didn't reflect anything like metal objects do. It never behaved like the object it mimicked. It was just an effect, a purely visual lacquer to help users understand that the Voice Memos app worked like a microphone. The entire interface worked like this so as to be approachable as possible. Notably, this philosophy extended to even the smallest elements of the ui. Buttons were styled to visually resemble a button by being convex and recessed or raised. Disabled items often had reduced treatments to make them look less interactive. All of this was made to work with lots of static bitmap images. The first signs of something more dramatic did begin to show. On iPad, some metal slider sheen could respond to the device's orientation. So as you deleting a note or email to not simply make it vanish off the screen, but pulled it into a recycling bin that went as far as to open its lid and close as the document got sucked in. I remember there was one that would like, shred things when you deleted it. They had a bunch of these fun ones.

You can imagine the Apple designers in that era being like, today I'm moving on from Apple. You know, I was the designer that made the lid open and the thing swoosh into the can.

It must have been so delightful to just be a designer and just like focus on just that one interaction for weeks, you know?

Yeah, that became a critique of Silicon Valley, was just like, you're working on something like the idea that, like, something like this doesn't matter in the grand scheme of things, but beautiful things matter.

They do.

And even a feature like this, especially in the Notes app of an iPhone, was something that was being used millions and millions of times a day and is worthwhile to make it beautiful.

And so there was pushback when they went to flat design. The flat age started in iOS 7, introduced an entirely new design language for iOS. Much, much was written on it at the time. And as with any dramatic change, the emotions in the community ran quite high. I'll leave my own opinions out of it mostly, but whichever way you feel about it, you can't deny it was a fundamental rethinking of the visual treatment of iOS. IOS 7 largely did away with the visual effects for suggesting interactivity. It went back to quite possibly the most primitive method of suggesting interactivity on a computer. Some buttons were nothing more than blue text on white background, like the Inbox button. The home screen, once a clear reference to the buttons on phones of yesteryear, was now much more flat looking. One part owing to simpler visual treatment, but also a distinct lack of shadows.

It is hard to look at this after looking at the last era. Oh, yeah, the Morphic era was amazing.

Goaded.

It was amazing.

It was amazing.

I see why they had to move on in some ways because it was just like the previous era felt it.

Started to feel dated. Yeah, it felt like.

It feels vintage. I mean, it looks vintage today.

Totally.

I'm sure the kids today will like, jailbreak their iPhones back to looking.

Yeah. I mean, I saw a comparison of the original Photos app, which looked like a camera lens, and then the flat design camera app had just a camera icon that was completely flat. And now we're back to a camera lens in the new era. And so some of that skeuomorphism seems to be coming back with the Glass redesign. Liquid Glass redesign, but not fully. You know, we're kind of splitting the difference. We're going in a different direction.

I just looked up. Do people still jailbreak iPhones? Yeah, and Reddit says it's still possible. The big thing is there's not much need for it. A lot of the things jailbreaking was for have been implemented into iOS.

Yeah. Because the App Store is so robust.

Yeah.

So you can get. You can get an app. There's an app for that now. There didn't used to be an app for it. There were many, many, many, many apps that were not available. And on the first iPhone, there was not even an App Store, so. But why did shadows have to go? They had. They had an important function in defining depth in the interface. After all, looking at the screenshot above actually does it no justice. The new iOS 7 home screen was anything but flat. The reason was that the shadows were static. IOS7 embraced the notion of a distinct visual layers and using adaptive or dynamic effects to distinguish depth and separation. Why render flat highlights and shadows that are unresponsive to the actual environment of the user when you can separate the icons by rendering them on a separate plane from the background? Parallax made the icons float distinctly above the wallpaper. The notification center sheet could simply be a frosted plane pane above the content, which blurred its background for context. Jony I've spoke proudly at the iOS7 introduction on how, quote, the simple act of setting a different wallpaper affected the appearance of many things. This was a very new thing. And so they're not losing the depth. They still. They're bringing the depth and they're actually making the depth more, more complicated with the Liquid Glass feature. And some people are frustrated by that. But one thing that they're, that they're highlighting is that it's intense from a computational perspective. Everyone was saying, like, rendering all these new features is definitely going to hurt your battery life because it's so much more. Although Apple is fantastic at integration and they're fantastic at this type of hardware and software integration and these types of renderings, they will probably be very quick. So I wouldn't expect a disaster, but I would expect a little bit more of a load just from the graphic layer, although it will be heavily optimized. Also, a new thing in the interface was that the UI Chrome was able to have the same dynamics. Things like the header and keyboard would show some of the content they obscured shining through as if it fashioned as if fashioned out of frosted glass. While it was arguably an overcorrection in some places, iOS7's radical changes were here to stay. With some of its dynamic effects getting greatly reduced. Parallax is now barely noticeable. Over time, its UI regained a lot more static effects. I wonder how much Parallax there is. I don't really notice any right now. I mean, I guess when you swipe around, you can kind of tell they're on a different plate. But I might have it turned off. I don't even know. More interface elements started to blend with the content through different types of blur, like the new progressive blur. And the button shapes were slowly starting to make a comeback. It settled into a stable state, but it was also somewhat stagnant. For bigger changes, there would have to be a rethink. What would come next couldn't simply be a static bitmap again. It would have to continue the trend of increasingly adaptive interfaces. So. So he goes on to talk about the age of physicality and vision OS and kind of where Apple's going and predicts a lot of what wound up happening. So Ben Thompson, to go back to him, he talks about Apple's retreat to empowering developers and partners. This was your point about maybe Apple will be auctioning some stuff off. We'll see how their service revenue grows over the next few years. But first, let's tell you about Vanta Automate Compliance, Manage risk and prove trust continuously Vanta's trust management platform takes the manual work out of your security and compliance process and replaces it with continuous automation, whether you're pursuing your first framework or managing a complex program.

It's like a AI golden retriever for mitigating risk.

Yes.

And automating compliance. It's literally a golden retriever turned into software. It's fantastic what the golden retriever does for your yard.

Yes.

Fanta does for your business.

Yes. Yeah. It's like having a golden retriever in the office. Like you just feel safer.

You do and you are safer.

And you are safer. So Ben Thompson writes about Apple's retreat to empowering developers and partners says that's not to say there weren't some notable AI announcement in announcements in Apple's keynote first Apple announced the Foundation Models framework. This is a quote from the Apple team. This year we're doing something new and we think it's going to be pretty big. We're opening up access for any app to tap directly into the on device large language model at the cloud core of Apple Intelligence with a new Foundation Models framework. This was what I was asking Alex about. Alex Heath from the Verge the question of hey there. Even if, even if the models aren't frontier, just the fact they're happening on device, the speed is very important for a lot of these models. Being able to use a model without a connection to the Internet. People are still, they're still missing Internet on planes occasionally or they have patchy service. Building an app that takes advantage of that and opening that up to the developer ecosystem where you can get really creative and kind of start finding those like niche cases of where AI can help for certain things. You know, we need the beer app, the Ibear app of the AI era. It's coming. We've been begging for it with this development. For example, if you're getting ready for an exam, an app like Kahoot can create a personalized quiz from your notes to make studying more engaging. And because it uses on device models, this happens without cloud API costs. So if you're this like kid developer who doesn't, who, who maybe doesn't have an economic model yet, you can build an AI product that only uses on device compute and you have zero. You have zero like cloud costs. Yeah, which is amazing.

I mean of course there's way less attention than it should have.

Yeah, I mean everyone was expecting it and it's great. But it is very cool in terms of just you can now Vibe code an iOS app that has some sort of AI functionality. It's not going to be incredible, it's not going to be frontier, but you can get something that's served into the store that only uses Apple's APIs and creates an experience that is, that is zero ongoing cost. It's entirely hosted by Apple and runs on the device, which is amazing. I think that that's very, very exciting to see where people go with this. They give another example of if you're camping off grid, pouring over the hikes you downloaded to all trails, just describe what mood you're in, the mood you're in and all trails can use the on device models to suggest the best option. So doing, doing things like recommendation algorithms on device, much, much faster, much easier. We couldn't be more excited about how developers can build on Apple intelligence. And so it's important not to oversell the capabilities of Apple's on device AI models. Of course, developers who want to create something that is competitive with the output of something like ChatGPT, we'll need to use cloud based AI APIs. Of course you got to burn, you got to burn up the H2 hundreds to get the really good stuff. But you know, it's like mid midwit. Midwit intelligence on too cheap to meter. Not quite, not quite super intelligence, too cheap to meter. But like you know, it's a midway.

We got that's helpful 2 digit IQ. Too cheap to meet really, it's bullish.

It's on the device. The reality. That reality however applies to Apple as well. Part of the folly of the initial Apple intelligence approach is that Apple was promising to deliver beyond state of the art capabilities on the cheap using its users, processors and power. What is compelling about the foundation models framework is how it empowers small developers to experiment with on device AI for free. This is what I was talking about, an app that wouldn't have AI at all for cost reasons now can. And if that output isn't competitive with cloud AI, then that's the developer's problem, not Apple's. At the same time, by enabling developers to experiment, Apple is the big beneficiary of those that discover how to do something that is only possible once you have an Apple device. Second, Apple deeper deepened its relation its reliance on OpenAI, incorporating ChatGPT's image generation capability capabilities into Image Playground and adding ChatGPT analysis to Visual intelligence. That's cool. There is still no sign of the long rumored Gemini integration or the ability to switch out CHAP GPT for the AI provider of your choice, but the general trend toward relying on partners who are actually good at building AI is a smart move. Third, Ben Thompson writes, Apple is incorporating ChatGPT much more deeply into Xcode, it's integrated development environment. This is their VS code competitor that you need to use to write objective C apps for the for if you're building iPhone software for building apps for Apple platforms, developers can also plug in other models using API keys. Xcode still has a long way to go to catch up to AI first ideas like Cursor, but again, partnering with a foundation model maker seems like a much smarter strategy than Apple trying to do everything itself. These are, to be sure, obvious moves, but that doesn't make them any less important, both in terms of Apple's future and also with regard to the theme of this article. Apple's initial success with the Apple 2 was because of third party developers, and developers were critical to making the iPhone a sustainable success. Trusting developers and relying on partners may be a retreat from Apple's increasing insistence on doing everything itself, but is a much is very much a welcome one. And so people are also, you know. He quotes Marquez Brownlee here. The Windows Vista update with Aero Glass was a huge part of my childhood, so I'm getting serious flashbacks. It does have a little bit of like the Windows Glass vibes. The biggest part of WWC was about Liquid Glass, which has drawn some compare unflattering comparisons to past Microsoft operating systems releases. Again, I'll withhold judgment until it ships, but there is another Microsoft OS comparison I've been thinking about recently. To me, last year's Siri disaster was a lot like Windows 8. Windows 8 was an attempt to micro to leverage Microsoft's large PC install base into a competitive position in touch based devices and it did not go well. Consumers, particularly enterprises, hated the new UI and developers weren't interested in the platform without users, Microsoft was forced to retreat and eventually came out with Windows 10, which was much more in line with traditional Windows releases. Apple has clearly missed the boat on cutting edge AI, but what I'm open to is the argument that this was a ship the company was never meant to board, at least when it comes to products like ChatGPT. Meanwhile, I've long been convinced that Apple has gone too far in its attempt to control everything even tangentially related to its devices. To that end, I understand why many people were underwhelmed by this WWDC particular, particularly in comparison to the AI extravaganza that was Google I O. I think it was one of the more encouraging Apple keynotes in a long time, and that's what Mark Gurman said as well. Apple is a company that went too far in many areas and needed to retreat. Focusing on things only Apple can do is a good thing. Empowering developers and depending on partners is a good thing. Giving even the appearance of thoughtful thinking with regards to Apple App Store is a good thing. Of course we want and are excited by tech companies promising the future. What is a prerequisite is delivering in the present, and it's a sign of progress that Apple retreated to nothing more than that. Let's hear for Ben Thompson. Another great article. Nailed it. Very, very good.

Hall of Famer anyway, future hall of.

Famer if you're looking for the Apple of Sales tax, get on Numeral Numeral HQ Sales tax and Autopilot spend less than five minutes per month on sales tax compliance. It's AGI sales tax baby.

Get on Numeral now.

So Sarah Guo at Conviction, former guest on the show friend of the show, said that Apple transcription is still so bad. So so bad compared to what's available on the market. They should be embarrassed. Okay, so I agree. I am now in the pattern of if I want to send you a long text with a lot of thoughts, I will go to ChatGPT. I will click the dictation button, I will talk to it and then I will copy and paste that into the imessage window.

Do you instruct Chat GPT to synthesize it?

I don't organize it, I just use it for the transcript. I just use it for the transcript. I open up the chat GP up, I let it transcribe. I don't even send the prompt.

One of those. John hits me with one of those yes, yeah bubbles. It's like this long. You actually don't do that. I don't really Notice.

You don't notice that I do it. But even if I'm just saying, like, okay, I'm walking around, I can't really stop.

And here's what, here's what we're hitting in the gym. You wake up, you wake up at 5:00am, first thing, transcribe the workout.

Five for five. 225 bench press. Yeah. So basically I agree it's bad. The question is, why is it bad? Because there are so many companies out there that they could work with, licensed by, partner with. There's a million different ways to solve this problem. Why is it bad? And I was thinking that maybe part of it is, let's say they buy a company that has amazing transcription and it is discovered that that company trained on data that it didn't fully have access to.

Interesting.

All of a sudden, instead of going after some startup with $100 million raised and 50 million in the bank and you're going to take them for all you're, for all they're worth, you're going after Apple.

Yeah.

And you can maybe get a billion dollar settlement. And so that's like a huge liability. And it's also potentially a PR nightmare because a lot of Apple customers care about the, the AI training. Adobe had to be very thoughtful about this in the sense that like, they're like, hey, look, we will let you opt in to let, letting us train our image, our models on your images. If you're on Behance or one of these other services. People don't like the idea of, of training just happening on their hard work generally. And so I think that there's some PR risk, there's some legal risk, but I agree that they need to figure it out.

Way less risk if they're just leveraging a vendor that is giving them even.

That, even that is like, it's tough for Apple because if they, if it's like, oh, well, they're working with this vendor and the vendor's not above board. Okay, well now they're, they're LinkedIn.

Yeah. But they should also be able to go to somebody like Scale AI or Merkore and just say like, hey, we want to build a transcription product and we're just going to make it ourselves.

I mean, they face, they like, they face, they face backlash when Foxconn has bad working conditions. Right. And so if they are partnering with another company and then there's some viral news story, whether or not it's real, that those people, I don't know.

But everything Apple does carries risk.

Yeah.

Even if they have great intentions and.

I don't need to figure it out. And this is something that just feels, it feels so easy to move the ball down the field because they already have a transcription product.

Yes, they just, let's just make it better.

They just need to make it better.

What are the ways in which we can make it better?

That's clearly not the modern transformer based architecture. Yes, they need to do a training run. Yes, they need to build some GPU data centers for it. Yes, they need new training data. But there has to be a way to do that in an environmentally friendly PR compliant like, like do it, do it the Apple way. I'm fine with that. But like you got to do it because this is such a basic functionality of a phone now. And the fact that I'm going from one app to the other just to just, just because the Siri button isn't the transcription is just not good. And that would be such an easy thing to just say hey, we're state of the art in this. Because this is something that is actually improves the daily experience of the phone user for sure.

And same thing with so backwards to have to go in another app, just send a message, transcribe a message.

Yeah, I mean I'm fine with them not inventing the newest paradigm of deep research. Like that doesn't. That product doesn't need to come from Apple. Yeah but we've had good dictation and transcription for years now and it's open source and it's not even, it's not even that. It's just like open source. I'm sure they can't just like vend the open source model in because of the license and I'm sure there's the restrictions on that. Like you're not just going to give whisper to somebody's company.

We should have somebody that's a big dictation head and, and have to fly to get a job. No, no, no. It's just somebody, just a fan go. You know, an Apple fan goes joins Apple purely to just, you know, beat the dictation drum internally because I feel like it just needs to be a handful of people that, that are really power users.

Anyway, I wanted to highlight this post from Pirate Wires. They did an interview on April 16th with David Hannamer Hansen. That's DHH of Rails fame. He also works at 37 Signals with a friend of the show Jason. Screen recordings of Apple's new iOS 26 Liquid Glass software design appear to show a jumbled, difficult to read apps and incoherent home screen app animations. For a company that built its brand on seemingly impossible feats of technology, Apple now finds itself stuck in a self perpetuating cycle, falling behind in the AI race, struggling to attract talent needed to catch up and fumbling product releases. Pirate Wires writer Blake Dodge diagnosed the company's problems in a May interview with 37signals. DHH. Read the interview in full on the site Link threaded. I recommend you go read it. I won't read it here. You have to go subscribe to Pirate Wires also. It has a lot of bad words but he really let loose and I'm sure there's a lot of insiders guarded.

State of affairs at Apple.

Yes, not quite family friendly material, but if you're an adult, head over, hit the subscribe button, check it out and read it in full because there's a lot of insight there. Anyway, that is the WWE Reset recap. We have seven minutes until our first guest. I thought we were going to do that in 15 minutes. Wound up being a 55, 53 minute segment. Love it.

Nice. Well, should we cover scale?

Yep. So meta is paying 15 billion. I thought it was 1415, something like that. For 49% AI that pays out investors and employees. CEO Alex Wang is leaving to join Meta and head up a super intelligence lab. Feels like those licensed and acquired deals at Character AI inflection and adept to avoid regulatory scrutiny, says Tane, who's been on the show. This is from the information you got to hit the size gong 15 times for Alex Wang. Let's hear it for a fantastic I just pulled out my your headphones.

I'm going to need to let you hit the size.

I will hit the the gong. Where's the gong mallet? Oh, it's over there.

Where's the mallet over there.

Here we go.

Technical difficulties, lots of difficulties.

But we got, we got it done. Congrats.

That was not 15 hits, John.

But that wasn't 15 hits, but it was one big one. And and it's very exciting. I I'm a big fan of Alex Wang. I've met him. I interviewed him on stage at an event about a year ago and he has a ton of really deep insight into artificial intelligence. And I'm sure he's going to be an absolute beast in the board.

Absolute dog.

Hopefully he can get a board seat.

A boardroom general.

He's a boardroom general.

Yeah. So it's funny when these kind of deals are relatively recent phenomena. Yes, we saw this with Character AI.

To Google to Google Adept to Amazon and Inflection to Microsoft and now Scalia to Meta.

And the interesting thing here, I was talking with an early scale AI investor this morning and they just still didn't, you know, I think the deal is still in the works. They didn't have full insight into how exactly it was structured. The question is Meta is paying $15 billion for 49% of scale AI and scale AI through that is losing their CEO to Meta. Now, scale AI is a very big business. They have a ton of customers that I imagine are going to want to continue to work with scale AI. The question becomes what happens to the actual EV of Scale AI in a scenario where half of it is owned by a hyperscaler who has very different incentive structure than a traditional acquirer. The positive here is that didn't you.

Talk to somebody at Character or Adapt or Inflection? You talked to one, somebody who was.

Yes, I talked to, I talked to. What happened with Character is basically like the cap table was effectively wiped out and reset and to my, not wiped.

Out, not cashed out to the tune of very high valuation. That's the thing.

I ended up in a situation where they had this massive balance sheet, no investors, like, no preferred shareholders. It was effectively is effectively, to my knowledge, community or sorry, employee owned. And I don't think that's what's happening here. But it is seemingly clear that the founders, the team and the investors in scale AI will be getting paid out, I would imagine pro rata based on that $15 billion mark. And the thing that's wild here is, I mean I.

So it's an up round for everyone. Everyone gets, everyone gets cashed out, gets a ton of money. Alex gets to go work in like a hyperscaler super intelligence and actually like take a run at some really serious like high technology.

Yeah, the thing that was ultimately surprising.

So scaly I has more than $900 million in cash on its balance sheet at the end of last year. According to financial information shared with prospective investors, they've raised more than 1.5 billion from investors. So they only burned 405,600 million of that. Like they kept more than half of it. The company made 870 million in revenue last year and expected more than 2 billion this year and it lost about 150 million. So I mean, doing pretty well. I mean there is a world where like Scale AI as a company, even though they're not wholly owned by Meta, they are independent and they have $900 million war chest and they can go build a great business and Do a bunch of stuff there and then that the relationship with Meta can be a fantastic growth engine.

Yeah. The only reason this was surprising to me. Well, I guess one it's Scale AI was if you're trying to build a new research lab, they're probably going to need to still acquire a lot of other top talent. You can imagine Meta going and trying to bring people from anthropic or OpenAI into this new organization, but it felt like Skale. I do wonder if Scale Scale's team could go back and see Circle's IPO and the success of that if they would have still done this deal because.

Scale AI could have ripped in the public markets.

Yeah. Circle IPO showed that the public markets are very receptive to pure play technology bets that have great narrative around them. And I could easily seeing what Circle has traded at. I could see Skale trading well beyond this valuation they're getting from Meta.

I agree with you. The question is how does the liquidity picture. What does the liquidity picture look like?

In that scenario, insiders wouldn't have been able to unload 15 billion.

Like does that. Does that actually play out? Because. Yes, yes, the, the stock could have traded up to 100 billion market cap. But can you actually, if you're an investor, can you actually get liquidity without crashing the stock if you're putting 15 billion of downward pressure? So, yeah, that's a lot. Big question.

We'll have to ultimately huge pickup for. For Zuck. I think Alex, you know, will. Will continue to fantastic, you know, execute very well and I can imagine he'll be a huge part of recruiting.

Yeah, I mean, he's done a fantastic job. He's like been everything. He's been on everything from like the Theo Vaughn podcast to testifying in front of Capitol Hill. And he's got range lawmakers. He's got a ton of range. He's very able to storytell and do creative deals which increasingly are extremely valuable in the world where the AI paradigms like the secrets are known. It's about, okay, how can you actually marshal the capital and the will to build the hyperscale data center to actually go and get Jensen to fork over 100,000 H2 hundreds on schedule? I'm sure there's a ton of different things in the works.

Yeah, it's interesting for Meta too. Even if they're able to leverage the Scale team to simply get better at ads, they can make back the 15 billion.

Oh, yeah. There's a ton of different ways this works out. And I mean they clearly need a lot more reinforcement learning training data for to improve llama 4 because they need a. They need a reasoning model if they're going to stay in the game. That's clearly the next paradigm. And I mean, like a big part of what we learned about from semianalysis this week was that verifiable rewards are really important. Reinforcement learning Scale AI pioneered that humanities last exam. They've been in, like the Evals game, defining the rewards that you would work against in the reinforcement learning paradigm. And so there's a whole bunch of ways that that plugging Alex and the scale team in makes a ton of sense. Anyway, really quickly, let me tell you about Adeo Customer Relationship Magic. Adeo is the AI native CRM that builds scales and grows your company to the next level.

Let's give it a go.

Check it out. And we have. Let's give it up for Adio Red.

Point, our first guest. Investors in Adio.

No, wait, really?

Yeah.

What a coincidence. One he washes the other. Welcome to the show. How are you doing?

What's going great, guys.

Great to be on. Huge fan of what you guys have built. So this is exciting.

Thanks so much.

Great to have you.

Would you mind kicking us off with a little bit of an introduction?

For sure. My name is Saison Lakumar. I'm a partner at redpoint Ventures. We're a venture capital firm in the Bay that's been around for over 25 years. We've invested in companies like Netflix, Stripe, but we're known for infrastructure software investing. So companies like Stripe, Twilio, Hashicorp, Kribble, Chainguard, and yeah, that's been our bread and butter.

And walk us through what's going on today.

Yeah, so we're in San Francisco at AWS's builder loft. We partnered with NASDAQ to actually move the their exchange to SF. So in about 30 minutes, we're going to ring the bell.

Really about.

In SF people here. Yeah, it's. It's like, it's like that Dalian tweet, like, move, move Silicon Valley to Miami. We just move Wall street to the bench here. So we're excited. Exciting day for us. But yeah, this is our annual conference and we do it every year.

That's amazing. Breakdown.

Hopefully you guys have. We have a gong cam here. Hopefully you guys have a bell cam set up.

But the gong might be better than the bell, I think.

Yeah, SF is very gong coded.

Well, yeah, I want to know about the top themes for the conference, the top tracks, the top discussion points. What are People debating for sure. Take us from kind of where the discourse is and where the frontier of that discourse is going, for sure.

You know, we created this whole initiative a few years ago to call out the importance of infrastructure software as standalone category. It really is like the picks and shovels behind how software is consumed and delivered. It's this invisible piece of software that we just take for granted. When you buy something on Amazon, for example, there's some application that we open, but it's powered in the back by all of these dev tools and databases. There's some global server taking in billions of bits of information and you need to secure it from all of these DDoS attacks, which happens millions of times a day. And we just don't know about that. So I guess across the whole application delivery lifecycle, there are just a ton of different infrared tooling and it powers virtually everything we do. So, yeah, we do this conference every year to just shine a light on all of the amazing innovation that's happening at the infrastructure layer. So we divide the world up into, into four different subcategories. Everything happening at the AI and model modeling layer, all of the data infrastructure tooling that needs to be built out around that, the cybersecurity vendors that need to protect that data, and all of the developer tools that actually creates the software. So we've been, we've been doing it for. This is our third year doing it. And we also highlight Basically the top 100 fastest growing private infrastructure companies. We all get them in a room. We have the CEO of aws, Matt Garman coming in chat today and the president of yc, Gary Tan. And so, yeah, it's a fun day.

I have this thesis I want to bounce off of you and get your feedback on. It feels like DevTools infrastructure. It's less monopolistic. And maybe that makes it easier for maybe not VCs to make money, but entrepreneurs to make money. You just don't get steamrolled as much. Am I off there in thinking that a lot of these dev tools markets are more oligopolistic by nature of the fact that hyperscalers are competitive and the different cloud platforms are competitive. What is the current dynamic? Because I feel like the narrative is always like someone comes out with some infrastructure or dev tool and it's kind of a boring company. It flies under the radar for six years and it gets. And then, oh, it's worth a couple.

Billion dollars and that's worth like 20, 30 billion dollars.

No one knows about it. Exactly. So we break down like what structurally is going on there? Why is it harder to build these crazy, crazy monopolies? And then is that narrative about like base hits being easier or unicorns being easier, is that even real?

I actually know, I agree with that sentiment maybe a couple of years ago, but if you look today, I mean look at cursor had $500 million in ARR in less than a year. And I guess like dev tools, it's always been this like very sequential, like if you think of DevOps, it's this very sequential process, right? It's like you have, if you have your engineers on one side, then you have your SRE folks on the other and then there are, there are tools for the developers, there are tools for the incident response engineers and there are like these distinct stopgaps that exist, exist across the life cycle, right? But with these co generation tools like Cursor and Windsurf and Copilot, like they all started on the left and they started in the IDE where these developers are actually coding but you can see the writing on the wall and where they're going. So now you're consuming and writing all these applications in the IDEs, but they're probably going to extend, right? And they're going to start doing the code review stuff and then you're going to deploy, deploy the apps. Like you can see it in lovable for example, with just like a simple query you can go from that query to a full blown production app that's live on a website. So like before you needed like 10 different designers and engineers to do all of that. But now with just one tool you can do that. So I think, I think yes, it's not been as monopolistic in the past, but I think that world is changing and like the clear horses right now in my opinion are Cursor and anthropic in the coding world.

Interesting.

And, and so yeah, I mean it's just crazy. Like two years ago like people were saying LLMs, all they could do is just spit out code. Like you couldn't actually write real software and impact software development. Q Two years later I think there was this interesting set like computer engineering grads face double the unemployment rate of art history majors and it's, it's, it's just crazy.

Seems crazy. How do you compare and contrast the adoption of, of AI versus traditional cloud? I know that you guys have done some, some research around, you know, cloud adoption, you know, more than a decade at this point versus even just the growth rate on the token side.

Yeah, I mean you Know, it's, it's just incredible how, how much faster the consumption is on the AI side. I think a really good analogy or data point is if you look at the cost of EC2 like servers during the cloud, it got efficient very, very quickly and it allowed people to consume cloud software all across the board. But when you look at the cost of inference, which I think is the equivalent to EC2 and cloud, but inference for AI, it's literally dropping like 100 times faster than the cost of EC2. But at the same time applications are being built and consumed 10 times higher than it was during cloud. So it's like 1000x more consumption at the end of the day. So it's just insane what's happening in the age of AI and the markets itself, there's this massive services component that is going to be converted to product based software revenue because you can encapsulate all of these workflows with software. Like LLMs are software at the end of the day and you're taking these huge markets and you're making them available through software.

How are you thinking about competitive dynamics between sort of scaled infrastructure providers like Databricks and Cloudflare and companies in that category versus some of the upstart companies that are scaling rapidly but you know, still having to compete with these, you know, founder led companies that are, you know, basically they're not going to let, they're not going to miss this sort of like platform shift.

Yeah, it's really interesting. I think the existing infrastructure vendors, they can actually embed the AI products in their suite. Take vector databases for example, which I think is an incredible market still. And there's all of these use cases. But is there a vector database equivalent of databricks yet? No, actually what happened was that MongoDB Extended Vector Search part of their, as part of their suite. So I think like, but, but like that's, that's not to say like, like for example, like two years ago people would not have thought an IDE like Cursor would, would be better than GitHub Copilot. So I think there are pockets of the infrastructure market where sometimes the incumbents are going to win and they can just leverage their existing technology to extend their applications. But then there are like AI native wedges like the IDE for example, where there is tremendous opportunity for a couple of 22 year olds to come and create a $10 billion business in two years.

Yeah, let's stay on Cursor Windsurf. The new, the new dev tools like IDE Coding AI coding market. I've kind of been. There's a little bit of a horse race going on. You know, you're either a cursor guy or a windsurf guy or you're an investor in one. But as I talk to more and more of the founders, I just increasingly become convinced that this is such a growing market and it's remains potentially oligopolistic that they might kind of all wind up winning in some way, or at least returning investor capital and making the founders fantastically wealthy and successful. How much of that narrative do you think is real? We talked to Scott Wu over Cognition. We were like, oh, like OpenAI just launched Kodaks. Like are you cooked? And he's like, well we grew 40% last month. And so it seemed like very much not so. And we talked to another person who was saying that GitHub copilot, not the trendy tool at all at a $500 million run rate. And so that in and of itself could be a public company if it was spun out. And so it feels like when we're talking about doing something so net new, something so additive to companies workflows, there's just so much opportunity that it kind of just is like a rising tide lifts all boats. But what do you think about that narrative generally?

Yeah, I mean if you just look at the application with the strongest product market fit within the AI world, it's coding and like the market itself, there are tens of millions of developers. If you just assume moderate assumptions and what you pay them, it's like a $1.6 trillion market in spend. It's absolutely massive. It's like the mother of all markets. And I think it's really interesting. There are companies going after certain pockets. So you have the co generators living in your idea and that's where the developers are today. Right. It's where the actual development happens. So cursor, it was just magic in a bottle. It's tap, tap, tap. And all of a sudden you have a full blown production app and you save yourself countless hours in development. But when you can see where the world is moving, like all of these processes, think about humans at the center. But it's like very, very quickly we're thinking it's going to shift to having agents at the center. And so companies like Cognition and Factory, they're not going to exist in your ide. Like you're just going to go tell it, you're going to go tell an agent to go do something and it's going to come back and have a pull request and you're just going to approve. So I think there will always be a need for software developers, but I think they're going to move increasingly from actually coding to being kind of these orchestrators of, of a product. And I think like product managers for example are now enabled with these tools. I think at the end of the day, like I don't, I don't think it's going to. Yes, there's like some developers that might lose their jobs, but you're actually just making their jobs a lot easier and there's just going to be a ton more software.

There's going to be more software.

If you're not Jevons Paradox pilled, now you never will.

You're lost.

Reacting to the news today, the scale AI Meta deal. How do you expect the data labeling market to evolve now that, now that in some ways one of the key players in that space has other. It will naturally have other priorities if you know half their company is owned by Meta who has their own goals around superintelligence.

Yeah, look, if you're the model providers, there are two use cases that are just so clear and hair on fire. The one that's like inferent is actually running these models. Two, it's data labeling. How do you actually structure your data and clean your data and feed it into these models such that you can actually advance the LLMs. And so I think it's a brilliant acquisition. I think whatever the sticker price is, we're going to look back, it's going to be like the Instagram deal. You need accurate classified data to feed these large language models and for a variety of use cases. There's so many data labeling businesses that we see like pockets in hiring and whatnot and coding and like they're all screaming out of the gates like they go to like 0 to 20 in like two months. And so I think, I think it just makes a ton of sense and I think it's, I think it's a brilliant acquisition.

I want to get your reaction to WWDC Felt like Apple was kind of pulling back from some of the territory that they'd over expanded into in artificial intelligence. Obviously it's developer conference. Do you think that there's opportunities for startups to build new companies even around servicing the nascent AI for Apple environment? You could imagine there is like there is no cursor for Xcode. There's probably a way to solve that problem even just anthropic.

And their deal with Xcode. Yeah, they're Working on it.

Yeah, but, but, but there's probably some way to even advance that. And then also just serving up models into, you know, iOS APIs that are like more tailored. Like there's been a couple companies that developed backends that were specifically tuned for iOS apps in the past. Some of those wound up getting sold. So how are you thinking about opportunities coming out of WDC or just general reactions? Maybe you don't even follow it at all.

You know it's interesting when you look at, when you look at like past platform shifts like the cloud you first need, like the workdays and servicenows and Salesforce is running at massive scale, like pushing the limits of like cloud rails before you can, before you can really understand what kind of innovation needs to happen at the infrastructure layer. So you know, our guest here with Apple and all of these apps that are built like you need the AI native versions of these apps that will start like hitting the limits of the underlying models before you can really understand. So for sure it's going to happen, but I think it's going to take some time.

What is, what's the discussion or what do you expect the discussion to be around agents? There's great narratives around agents right now. There's individual agent products that people use, whether it's coding agents or sort of research agents. But despite, you know, Salesforce, you know, marketing, whatever their agent, cloud agent force, agent force, things like that, it doesn't feel like they maybe have broad adoption yet, but a lot of promise. So how do you think about kind of agent adoption given there's people, I think at the event on the browser infrastructure side, AI agent DevOps, all that kind of thing.

Yeah, look, I mean it's clear it's the next paradigm. I think we're just catching it in a point in time where they're stumbling agents. But again there are some clear use cases where it's clearly working today. Cognitions agents, for example, there are low level development tasks where that is fully automated and the largest of companies know that, but no one else really knows that yet. And it's mind blowing what's happening. Same thing with customer support, your companies like Sierra and Decagon. And that's just a use case that makes a ton of sense for agents to operate. So I think as the agent infrastructure improves there are things like memory and context budgeting. There are a lot of these like hairy infrastructure problems that need to be solved but ultimately also the agents are going to improve dramatically like from a year ago, you know, agents have Just gotten so much better. A year from now, they're going to be even better. So as agents improve, all of the infrastructure around, it improves, like memory, like. I think very, very soon there will be an agent for every industry.

Do you think it's. Do you think that it ends up looking something like regular SaaS or from A. From a. From a user standpoint in terms of you have sort of a dashboard, or do you think there's potentially totally novel paradigms?

Yeah, you know, I think UI is, is just a big problem or just an opportunity for these companies to kind of rethink how you surface these products and extend these products to your end consumers. If you look at agents running in the developing world, in the development world, they're just running in your cli in the background. Is that the best way to do it? I don't know. Maybe voice is the best way to interact with them in the future. I think that UI is still being reimagined right now. I'm guessing what we have right now will be thrown out the window and that these abstraction layers will be very, very different in, In a few years.

Yeah, the voice thing is interesting. That was actually in that original Vibe coding post from Andre Karpathy. He says he doesn't even touch the keyboard. He just uses Whisper or some sort of like super whisper app, he said. And he just talks to it and just says, yeah, just the padding. Accept all. You know, oh, there's a bug. Fix it. And yeah, I mean, it would be very. It'd be super, super weird because for the longest time, the programmer was, you know, the keyboard was sacred. It was. Don't have. You don't even have a mouse. Everything you do is on the keyboard. You probably have an ergonomic keyboard. You're dealing with carpal tunnel because you're on the keyboard so much. And if we move to a world where you still have a genius programmer who can understand the architecture, understand what questions to ask, what features to build, that type of stuff, but the actual interaction day to day is via voice, that would be a very, very big shift. Be very interesting. Anyway, I want to get your reaction to the news today that glean raised $150 million Series F at a $7.2 billion valuation.

Now, you said glean.

Glean, yes. So Glean. Their whole goal is to enterprise search. And I'm interested in particularly who are the beneficiaries up and down the stack, because I imagine that they're in a position to kind of say, hey, they stuff all your data in Glean, but they probably want to be more of an integrator and sit on top of your data lake, on top of your snowflake or on top of your AWS installation and plug into your Slack. How are you thinking about having tools that really allow folks in the enterprise to interface with all of the infrastructure that you have described?

Totally. And you know what's interesting is the category isn't new like enterprise search and there were so many startups and there's this graveyard of companies that existed. But I think the clear why now? Or at least the enabling technology was lm. So of course Glean is going to benefit. But like what's powering all of this are OpenAI's and anthropics like really, really powerful models. Now there needs to be this layer of infrastructure that exists between the model providers and Glean itself. It's like the hairy problems like memory and caching and so those providers are going to win as well. But ultimately the consumers also and these companies when it's. It's. It's amazing even for me. Like if I'm trying to search something in Google Drive, like how long it takes. Oh yeah, something, it's great.

It's degraded somehow. It's gotten worse. Like I can't even see my email. It's just like there's so many like cookies now that if I search for like Jordy, it'll find that in some random URL string and be giving me some like, you know, spam email I got. It's a mess.

Yeah. But you know, we should expect. And I think it'll get better. Not. I think it's OpenAI is going to release a product here. You will have like why does there need to be an application on top of this?

Oh yeah, totally. I mean they were, they were teasing this a little bit with the. What, what was it? The, the new. The new.

I think there was a little, a little bit of drama around different people not wanting other people to invest in Glean.

That was one thing.

Open AI. Yep, yep.

But then there was also the, there was also the latest news that that Deep Research now can search across GitHub, Google Docs, Gmail, Google Calendar, SharePoint. And so that's a Glean competitor. That's a Glean competitor.

That's a Glean competitor. They're just not saying it and so.

Exactly.

And so I think for the applications that make that there is strong, really strong product market fit with the large language models, we should expect the model providers themselves to move up the stack and address parts of these apps. Because, I mean, if you think about like the models themselves, like, yes, they make a ton of money, but they also burn a lot of cash because OpenAI is just trying to build that next frontier model. So they're going to go where the consumers are like, let me go buy Windsurf. Right. And be in the ide. Let me go release something in the enterprise search space to compete with vendors like Glean. But I mean, I think it's fascinating.

Yeah.

When are you ultimately. Or the CEOs that you're meeting with today? When are you. You know, it doesn't. I think there's areas in which people feel safe.

Right.

If they're working on like a specific application of LLMs with a bunch of deep integrations and workflows, people generally feel safe today. But is there. Are you expecting a lot of dialogue around people just trying to assess, like, am I on the lab's roadmap and really trying to figure out, like, at what point they're going to get, you know, Sherlocked?

Yeah. You either have to be. You have to be building on the right side of AI. So if you're trying to solve for deficiencies in the model today and you're trying to build these things very, very quickly in a couple of months, all of that tooling just didn't need to exist. So you just need to be building on the right side of AI. As these LLMs become more powerful, your application itself should become a lot more powerful. And how do you embed into workflows and also build all of the little hairy infrastructure things to empower all of that?

Yeah, it's interesting, ironically, like wrappers, aggregation, aggregating demand being a front door to AI. Like, that's where a lot of the value has accrued. Or at least like it feels like the foundation model companies feel most threatened by those approaches versus the threat or.

Just want to own them.

They want to own them or they want to compete with them directly as opposed to the. Okay, yes, you leapfrogged us on capabilities. That's much less of a threat then. Oh yeah, you beat us on a benchmark that matters a lot less than when people think of this particular AI use case they go to this website instead of mine. That's more threatening to the foundation model labs, in my opinion.

Totally. A good analogy is that the hyperscalers like GCP, AWS and Azure are the equivalent of OpenAI and anthropic for the AI world. And if you saw what AWS did For example, right. With Elastic they open source Elastic was a very permissive license and they open sourced it and they started competing with Elastic itself. But Elastic ended up doing was they, they went for a much more restrictive license but you still had AWS selling Elastics core product. And then Elastic is also existing as whatever it is a $10 billion business today. So I think going back to your point, I mean like it's, I don't think it's winner take all. I think the model providers are going to have a large portion of revenue and spend here, but they're also going to be really exciting AI native companies that exist on top.

Yeah. Yesterday we talked to Vlad from in physical working on API keys and secret management.

Really cool company.

Yeah. Raised a $16 million Series A from Elad Gil. I wanted to know more about different strategies and approaches to building infrastructure plays that either play directly or alongside open source strategies. And so there's an open source package out there that's gaining traction. How do you build a product or company around that? Or you have a company that's built something, do you open source it? What are the different strategies that you're seeing, what's working, what's gone out of fashion? I just love some color on how to build either with or alongside open source in the enterprise.

Yeah, I think open source is a double edged sword. I mean obviously we've been fortunate to back companies like Hashicorp and Clickhouse and we've seen the commercial success of those businesses but ultimately for these companies the biggest competitor is actually your open source project. Right. Why do I need to go pay you millions of dollars when I can just try to do this myself and fork it. And so for the open source companies out there, I think you need to be really careful and just kind of understand where your product is going, the types of enterprise features that you could add that, that buyers would actually want you to pay for. Ali at Databricks, he has a good analogy here. Like building a massive open source business is like hitting two home runs simultaneously. One, you need a really exciting product where there's a ton of open source traction and two, you need to figure out monetization in the case of databricks. You know they had Spark and it was great but they had to figure out a way to make spark really, really cheap and easy to use. And, and, and, but they had to hit these two massive home runs to build a big business. So for the open source founders out there, I would, I would say like really consider how you're going to commercialize this product.

Yeah. What, what are the, what are the case studies that people are coming back to today? I remember learning the, the history of Red Hat Linux. Built on top of Linux of course, fantastic business. Wound up of a couple public company. Then we've also heard the story of GitLab that went through Y Combinator became a fantastic business. Databricks has a similar story. What other stories are people latching on to these days? What are kind of like the, the, the, the, the, the famous examples that people keep pulling from in this open source enterprise infrastructure world?

Yeah, I mean MongoDB is one. Right. Look at the success of that business. And what's really interesting is that they released their cloud product much later from the actual open source. It actually came right before the IPO and it's actually accelerated their revenue growth since the only company where that's the case, the enterprise software.

So how are they making money before having a cloud offering? Just.

Well, they had an enterprise core version of their open source, but like got it. Then they also released a cloud self hosted version of that product. Hashicorp is another example. Right. With Vault and all of these systems. And so they're obviously like companies that have been successful in doing this, but it's, you need to be really careful and you need to think about monetization.

Yeah. It just seems so dangerous because you're building on top of an open source product. You know this is going to be baked into Amazon and Google and Azure like very quickly. But at the same time you're maybe a founder mode company. You have a lot of really aggressive people that can go and move and launch new features that can just keep you cut above. And then yeah, you know, the company says yeah, I could get, you know, 80% of what I want on AWS, but I want it to be perfect. And so I'm going with you because you're the best vendor.

Yeah. Basically is the total cost of ownership of having all of these developers internally and trying to maintain this in house, is it much less than the actual service that they're providing the enterprise service? If that's the case, then great. If not, then you really have to reconsider this, the strategy.

Yeah. Nothing's worse than having an open source installation and having to deal with like, oh, we got to turn it off and turn it back on. It's much better to pay someone else to do that.

And then, and then, you know, it just creates just really interesting dynamics with the open source community.

Right.

If I'm only going to release features in the paid enterprise version and I'm going to kind of just stop releasing features to the open source version. Then your community is going to get upset and so like. Yeah, you just have to be careful.

Yeah. Well, this has been a fantastic conversation. Thanks so much for taking the time during the busy day to check in with us.

Fun ringing the bell. Come back on again soon.

Yes.

Great.

So this is great. We'd love to see it. Enjoy the rest of your day. Yeah, big one. See you soon.

See ya.

Bye. Whether you're looking to get exposure to MongoDB or Red Hat Linux, go to public.com investing for those who take it seriously. They got multi asset investing industry leading yields. They're trusted by millions.

Oh, join interesting new campaign today.

How did you sleep last night? Because I went on a whirlwind tour. I fell asleep in my son's bed so I got to pick up. It doesn't count. I need to sleep in his bed and then I can have Tyler vibe code up something that merges the two data sets together because I think I got like nine hours yesterday. I'm feeling fantastic but I only log seven hours on my actual eight. Sleep, sleep. So I got an 86. How'd you do?

I got an 83. That's two nights.

Oh, two nights in a row. Let's hear it for John Coogan, the sleep master. No one out sleeps me. No one out sleeps me. Don't even try. Oh, we got some breaking news from the printer.

Definitely ill. We got some.

I'm gonna be a hand model now this is. This can be bad. Okay, we got some breaking news. Anduril has been placed number one on the CNBC disruptors list. Congratulations to Anderil. And Trey Stevens is jumping in on the timeline, taking a shot at Sam Altman saying ha. Take that, Sam Altman. Gotcha Crying emoji. Of course their buddies and founders funds and investor in OpenAI. So they're having some fun. It was crazy that Andrew got number one because OpenAI has what, 10 times the market cap at this point. But both fantastic companies. You'll love to see them at the top of the list. Number one. And number two, we should go, we should go through the disruptors list because we have a little bit of time before our next guest, Karina from linear is joining at 12:45. We have 15 minutes. Let's go through some timeline reaction. Let's go through some CNBC disruptors list. I want to see who's going on. So anyway, to recap the glean news they raised $150 million adding billions to their valuation. Their $77.2 billion valuation now very 7.2, 7.2. And it was a who's who on the cap table. The CEO Arvind Jain really using the long post feature on X, he gives shout outs to Wellington Khosla, Bicycle VC, Geodesic Capital, Archerman Cap, Elkap Capital One Ventures, City Ventures, CO2, General Catalyst DST, Iconic Institutional Venture Partners, Kleiner Latitude Capital, Lightspeed Suppliers, Sequoia. Everybody's in.

This tells me that OpenAI didn't really get their way and not wanting to people to.

Yes. So what you are referring to is the fact that back In October of 2024 last year, there's an article in Reuters that says OpenAI asks investors to avoid five AI startups including Ilya Sutzkever SSI sources say so as as global investors such as Thrive Capital and Tiger Global invest 6.6 billion in OpenAI, the ChatGPT maker sought a commitment beyond just capital. They wanted investors to refrain from funding five companies they perceive as close competitors. The list includes Anthropic, X AI, Ilya Sutskeverse, New company Safe Superintelligence and then where was the other 1? 2 app 2 application firms, perplexity and Glean. And so this is kind of a funny moment because you know, Anthropic seems like the direct competitor Glean was kind of just thrown in there.

But in many ways, but it makes sense. I don't know. In many ways Anthropic is not trying to compete in the core. You know, they're so AGI pilled, they're not trying to compete.

They don't have a, they don't have a voice model, they don't have an image model. They're just focused on that. So but at the same time they are competitive and you know, if you can keep your investors on your side of the table. There used to be like an unwritten rule in Silicon Valley that you only back one company.

OpenAI has just gotten so big, big. They're effectively, you know, public company scale.

We are the market.

So it's a hard ask to say yes. Just don't invest in SaaS Enterprise, other labs. Just don't invest.

Just don't invest.

Take a year off. Take a year off.

Take a long summer break, get stands.

Take a six month spring summer break.

Just put your whole team on getting into the massively oversubscribed round over there. Focus on that.

Good luck.

Work on yourself.

Work. Hit the focus on yourself.

Hit the slopes.

No more investing. Focus on yourself.

Yeah, but this, this news happened last week. OpenAI said we're rolling out ChatGPT Record mode to teams users on macOS capture any meeting, brainstorm or voice note. ChatGPT will transcribe it, pull out the key points and turn it into follow up plans or even code. This was taken as a granola attack, but it also plays into the glean story. So Signal said Granola is more of a social company than a B2B company. It punctured a norm. No big player would have taken the risk of recording without permission. But Granola beautifully reframed it as productivity intimacy. Even once the social acceptability shifted, it was inevitable that infra players like OpenAI and Notion would mimic the mechanic under the same pretense. This is the difference between product innovation and cultural mutation. Tech can't move where the culture hasn't first been softened.

And so, so interesting because Otter. Do you remember Otter AI? Have you used them?

I've used Fireflies.

Otter was started in 2016.

Yeah.

They. They crossed 100 million ARR.

Yes.

This year.

Yes.

And in many ways it's very similar to Granola.

I believe I inspired some of these because I don't want to take too much credit because I didn't build anything, but in 2014 I had a workflow where I would record, I would record meetings audio. I record the audio of Zoom meetings or, or Google Hangout meetings and then I would send those to a transcription service that was not using AI at the time and I would use Zapier to automatically do this or Zapier. And the guy and the team at Zapier reached out and they wrote a blog post about my workflow as being like a cool example of how to use Zapier in an interesting way. And a couple founders reached out to me and said hey, we want to build this as a product. And one of them was Fireflies. And it's like a banger company now. And obviously it's massively, massively enabled by AI and only gets better and better. But yes. Kevin weil over at OpenAI, the chief product Officer said Deep Research can Now search across GitHub, Google Docs, Dropbox, HubSpot, Outlook, Box and even more. You can connect any chat to Google Docs, SharePoint, Dropbox and Box. There's even an initial version of McP support for HubSpot.

This is big.

There's record mode in ChatGPT single sign on for ChatGPT teams. So they are firmly going into the Enterprise. And this was, this was a point that people were making, was that ChatGPT for teams didn't feel like an enterprise product. It just allowed you to provision a normal ChatGPT instance to an employee and say, hey, we're covering the cost. Basically, it wasn't truly integrated. Imagine how cool it'll be if we add someone to the. To the TVPN team and they, and they immediately get access in their ChatGPT to all of our internal documents, everything that we have about all of our transcripts, all of our. All of our.

Almost like an enterprise search product.

Yes, exactly.

Somewhat like glean.

Somewhat like glean. And so it was clearly on the roadmap and they, and they work towards it anyway. They're doing great. OpenAI's number two on the CNBC Disruptor 50. Then you got Databricks. Anthropic at four, Canva at five, Ramp at six. Let's hear it for Ramp. Congratulations.

Using Ramp. Love to see it. No surprises.

Safety at 7. We got to get the founder on the show. Very interesting company.

Crime fighters.

Crime fighters. They solve a ton of crime in America. More important than ever.

Alpha Sense, kind of like productized. Gary Tan.

Yes. Gary Tan is a service. Alpha Sense is at number eight. They of course, acquired Tegus and are a market research team. Team market research firm. Juggernaut Intelligence. Juggernaut, big advertiser on invest. Like the best. We love to see it.

How I was about to say, I was about to say, if there's a single company that I don't know, I'm cooked.

Yes.

I'm cooked.

Yes. But number nine, Octopus Energy. Look it up. Give us a breakdown.

Haven't heard of them. Smarter, greener, fairer energy for every Texas home. Is this a base power competitor?

Interesting. Is.

I don't. I don't know.

Do they make batteries or do they make solar panels? Break it down for me. Reading while you do some research.

So I'm on the website and I still don't know how it works.

Okay, well, all you need to know is that they're number nine. Let's have the founder on the show. Have him explain it to us. We don't need landing pages, Jordy. We get it straight from the source on this show.

Yeah, there's a bunch of graphics of Octopus. I don't know.

Wait, really? Is it. Is it energy generation, energy storage? You do some research. I'm going to read through some more. They got Stripe at 10, Revolut at 11, Thrive Market at 12, Metropolis at 13. We have the Founder of Metropolis. Coming on the show in just a few minutes.

Would you look at that.

Trans. Transcarrent. I haven't heard of this. They care. They're at number 14. We got to look up them. Lead bank. Fintech Fixer at 15. Haven't heard of them. Carbon Robotics, Weed Tech. We talked to the founder of Carbon Robotics.

That was awesome place for health and care.

Virta. Virta Health. What comes after GLP1? Statistics. Interesting company. Want to dig into them? Fruitist, the berry unicorn. We got to look into Brutus. That sounds awesome. Saronic at 19. We know them. Grub market is.

So what is the criteria for writer getting on this list?

We will look, we will dig into that because let's.

It's definitely not revenue.

There's a whole bunch of great companies on there. You got Figma, Zipline, Sierra Runway, Navon, Shield.

Fruitist is inspiring and inspiring, enjoyable and nutritious snacking.

Okay.

Fruitist grows the world's most delicious berries bursting with flavor, crunchy goodness and freshness in every bite.

Okay, here's how they chose the list. So they're all private.

Any of these?

Yes.

I love. I'll go out and say I'm a huge fan of fruit. I think it's one of. One of it's fantastic. God's greatest inventions and gifts. But have any of these fruit brands broken out?

Frutus. Seems like they've broken out on the list.

Beautiful website. Beautiful website.

Do you want to hear about how they chose the 2025 Disruptor 50? Okay, I got to know John. So they're all private, independently owned startups founded after January 1, 2010. They were eligible to be nominated. Companies nominated were required to submit detailed analysis including key quantitative and qualitative information. Quantitative metrics include company submitted data on their sales, number of users, employee growth or lack thereof, and more. Some of this information has been kept off the record as it was used for scoring purposes only. CNBC also brought in data from PitchBook and Ibis World and compared those based on the industries that they are attempting to disrupt. There's a board of leading thinkers in the field of innovation and entrepreneurship from around the world, along with a newer advisory board that ranked quantitative criteria by importance and ability to disrupt established industries and public companies. This year, the two advisory boards found that scalability and user growth were the most important criteria, followed by sales growth and access to capital and community. It's very funny that Andrew is number one because like, what are their user growth metrics? Like that's not the metric for them at all like users are a relevant metro. They're not looking at like mouse and daos. They're looking at like how many things did we build?

But still coming in. I'm still fixated on because I guess going off of users, you know they're selling consumable good and in whole foods. But tough, tough go for you know, companies like Scale AI which are down at 2028.

So Fruitist is better than Fruit is 18.

So yeah, if you were going off of this, you go short, possibly long fruit possible. This is should be meta's next target. Maybe they should try to move up the list.

Yes. And get them all. They already have a deal with Andoril.

They have an amazing distribution engine.

Yeah.

Every other ad could just be an ad for berries or fruit and I eat these things all the time and it'd just be a good way to really. They could potentially turn fruit is from the bear unicorn into a, you know, a deca unicorn.

I love it.

Deca corn.

We need, we need more food. We need better food. This is a common, common theme. This makes a ton of sense. So new for 2025. We can compare the way the two different advisory boards can considered the importance of the list criteria. While the two boards mostly agreed, the VC group thought that the size of the industry being disrupted was more important than the academics did, with the latter ranking access to capital and community as more important criterion than the group that provides said access. The ranking model is complex enough to be sensitive to these differences of opinion, and perhaps more than ever, it makes good on the concept that companies must score highly on a wide range of criteria to make the final list. Nominated companies were also asked to submit important qualitative information about themselves, including descriptions of the core business model, ideal customers and recent company milestones. A team of CNBC editorial staff, including TV anchors, reporters and producers, and CNBC.com reporters and editors, along with many members of the advisory board, read the submissions and provided holistic qualitative assessments of each company. In addition, the VC Advisory Board assessed a small group of finalists as an additional component of the qualitative review. Specifically, we asked vc the VC Group to assess some of the companies that would, if selected, be making the list for the first time, as well as to help in the consideration of high scoring early stage firms, a group with.

Lower value part of the criteria. They didn't want the top 25 to all be enterprise SaaS companies because that doesn't. That's not good for ratings.

It's not good for ratings.

Not good for ratings.

So only 11 of the 2025 honorees are pre chatgpt CNBC disruptors. But most of that group anderal databricks and Canva chief among them. The embrace of the new era is what has kept them here. And so there were 19 first timers in either 2023 or 2024. Companies are moving up, they're moving down. But it's an exciting list. Who else sticks out on this list? We got to get some of these folks on a better ride to school. From Zoom Wabi. How self driving trucks. See and learn. That seems interesting. Apptronic. 5ft 8 inches. 160 pounds, all robot. Haven't seen them. Harvey's on the list. Lawyering up AI. They got good pithy phrases, good lines.

Notion Project. Project management. Okay, we got project management for your projects.

That's kind of an odd one.

Not the strongest.

A bridge. Getting a doctor's note. Isesu Isusu Treating renters like owners shield AI.

We have the CEO on interesting.

Well, very fun list. We love a good list. Now that Conrad's out, the Midas list has fallen from grace and we need new lists. But congrats to anderall, congrats to OpenAI and congrats to everyone that made the list of the CNBC disruptive 51 things clear.

You got to be an absolute dog to get on this list.

Yes, you do. Well, we have Kari from LINEAR coming in the studio.

Get the gong ready. John, why don't you hit it?

I will hit it for him.

We're gonna hit it a few times. Kari, before you. Before you dive in. Yeah. Give us the numbers first.

How much did you raise? How much did you raise?

82 million.

Low dilution round at a billion.

Congratulations.

See it. We love to see it. A lot of people would have said, you know, this was maybe an unlikely round. You guys have been very, very efficient. But it's a testament to.

To proving the haters wrong. They said it couldn't be done.

Well, they said it could be done. They said K would never would wouldn't do it. It's great to have you on. Massive milestone for you and the team. How are you feeling?

Yeah, feeling great. And yeah, I think it's definitely one of this. I think all of our rounds has been kind of interesting because since I think the seed round, we haven't really needed the money. So each round is kind of like we're doing something different or we are signaling something to the market. So it's a little bit always strange to do these Rounds, because I think often you celebrate it as a way to be able to do something more. And I think in the end we will do something more. But also that not a lot of changes. We still keep building for the customers and the things we're doing.

Talk about maybe some of the last rounds, how they were unique and what makes this round unique.

Yeah, I think we also done this interesting thing where we have doubled down on each of the investors. So Sequoia led our seed round back in 2020, and then we did a Series A with them, I think 21, and then we did a Series B with Excel in 22, and then now two years later, or like about three years later, we did a Series C with them. So I think there's also, like, I remember when I was talking to back in the day, like around the siege round to a lot of seed round investors, and I think they had this. They create this fear that these platform funds can be challenging and there can be this signaling risk and all kinds of things. And I don't know if that's really true or I don't even know when that has happened, but I do know that it's kind of interesting. When you are working with these platform funds, they can double lead two rounds in a sequence. And I think the benefit there is that you can kind of push down the delusion because they already have some kind of ownership. And I think a lot of times with this game of fundraising, the actual absolute numbers don't matter how much you raise or what the valuation is. It's more like everyone else is fighting for their percentages. And so if you kind of work with the same investors, you have more, I don't know, leverage, or there's a little more, like, flexibility on those percentages.

Yeah, yeah, yeah. I mean, it makes sense. It also makes board construction easier because you're not adding 25 different board members from every fund. On the platform fund thing. I mean, it's clear that, like, that narrative came out of, like, kind of the clubhouse era, where the company got marked up multiple times by the same insider and the metrics didn't really meet that. Was this more of a Vibe round or an Excel round? How are the metrics tracking? What.

What are Excel round?

What are people watching? Are they looking like. Is Mao Dao the key KPI? You don't have to give us the actual number, but, like, what is most important? Is it revenue? Profit? Ebitda? Ebit? Like, what. What metrics are, are most important for a round like this to get done in this era, I mean like, I.

Think like all the later stage funds, like all later stage rounds, it is about the growth and the revenue. I mean obviously different companies are different, so there might be other things. But for us, since we sell the companies, it generally turns into revenue. We don't have a lot of free users or free customers. Everyone is a paid customer. I think it comes down to one is the growth and the position we have in the market. We being able to capture quite a lot of the growth market these days, like the early stage startup, but also like the late stage growth market like companies like Ramp or Mercury or Brex or, or and even open AI which is, I don't know if you categorize it as a growth company or like enterprise or. What is that? I don't know.

Behemoth company.

Behemoth. Yeah.

Yeah.

We need a new word, I think to corn something like that.

Yeah, yeah. I think like we, I think it comes down to like we work with the best companies out there.

Yeah.

And I think we also have a lot of people really love the product and they really want to use it. So I think we have had a lot of opportunities that we could have additional investors join the board or lead these rounds. But for me at this point, I think it's probably the next time around. There's probably reasons to find someone new and find someone with some kind of specific or different skill set, but right now I didn't feel like that's necessary. Right now I think like the biggest thing, what we're aiming for is that we have this core business and product going with a lot of like 15,000 customers like companies that I think that the next thing is the kind of like the AI and how, but you probably heard it this like a million times on this, on the show, but I think again the AI is shifting the market and how things work and how software is built. And I think often I like to do these rounds when we don't need to do it and we also. When we are about to hit something big. And I think we are about to hit something big with this.

Yeah. Talk about some of the discussions around talk about some of the discussions with Excel and, and some of the other partners and internally with the team around this sort of inflection point with agents. Because for the first, you know, you guys have had a bunch of integrations for a long time, but this is sort of a novel type of integration. I know you guys are working with, you know, Devin and Cognition and other players and you know that, that Feels to me like a lot of the linear's position as a platform that can help manage agents feels like, you know, a major catalyst.

Yeah, yeah. So some of those companies we've been working now but. And there's a lot more coming and I don't have the exact dates yet, but I think we're also working with the bigger companies out there that are building some of the like the, like the model companies. So hopefully those are coming soon. So yeah, I think like with the agents the idea is that we have this, we have built this like end to end workflow for like discovering and planning and building products. And it's kind of familiar to these organizations and these product teams. And I think right now we started seeing distraction from our customer base and also talking to CTOs. Almost every CTO out there is thinking about AI. I think the CTOs and the CEOs are writing these memos and they know that I can help their organizations. But the problem now is there's a lot of friction. The solutions are not that designed yet. It's a little bit hard to use some of these tools or use this technology. So the idea here is that we have this familiar system that people already use to do their work and so we can easily introduce these agents and some of the other AI capabilities into it. So they don't have to go somewhere else to do it, but they can continue what they're doing. But now instead of assigning tasks or code bugs or something to humans or the teammates, you can assign them to agents and then work with them to solve them. I think the idea there is that we want to make it as easy as possible for organizations to kind of adopt the AI or adopt these agents or adopt like the new technology because like a lot of the other companies are kind of like more building the underlying technology. But then sometimes it seems like it's hard to, for enterprises.

Orchestrate.

Adopt.

Yeah, orchestrate them.

I want to let you get back to building, but I want to get WWC reactions.

Yeah, we got to get your take on wwc.

I mean Linear is a fantastic design driven firm. I'm sure that there was a lot of internal chatter in the slack black or in the, in the, in the, in the group chats about wwc. How's the reaction?

Liquid glass.

Liquid glass. What's the take? Give. Break it down for us.

Yeah, I don't want to go too harsh on it until, and then I think until it's like in production. I do think like there's some discussion, I think that the borders are too rounded. And I think, like, some of the borders are off. Like, they're not following the right. Like when you have, like, two rounded rectangles, you just kind of like, follow the same curve with those round, like the border radius. So it looks like kind of like an even. Like that. There's even spacing alongside all of the kind of path. But I don't know if that was a mistake someone did for the keynote or they just didn't want to, like, I don't know. They didn't do it correctly. I hope it was a mistake. I think it's. It seems like really, I think, obviously visually interesting, but I'm still skeptical until I see it, how well it actually works in practice at linear. Actually, we did have some of this. We had this transluent UI at some point, and then I decided to remove it because I thought it was too distracting and actually it was kind of slowing the application down, so I decided to remove it. And so I'm hoping that Apple can make it performant and not distracted.

Yeah. Do you think there's always going to be readability challenges with it, or is it something that they can solve?

I mean, I think the starting point. It is hard to solve that because I think that was my experience with having transparent UI as well, is that it can work well if you have a nice, I don't know, solid color background on your computer or something. But then if you have a photo on the phone, you have a photo of your family or something, it can be kind of pride, and there can be a lot of different things going on in the background. So I do think it will be hard to make it kind of easy, the contrast to be enough to read it. But I don't know if they're gonna do some kind of magic there that they could figure out, like, what's on the background and then adjust it somehow. So I'm hoping that they can figure it out, but I remain a little bit skeptical until, like, actually, our batteries.

Our batteries are gonna be cooked.

Batteries are gonna be cooked. I think they'll figure it out.

Well, congratulations to you and the whole team and, yeah, excited to see what you guys continue to cook up.

Yeah, this is fantastic.

Getting started.

We'll talk to you soon. Have a good one.

Thanks.

Bye.

Have a good one.

Kari, let's check in with Tyler. Apparently he has iOS 26 installed. Give us the update. Tyler, how's it going?

Let's see it.

Yeah, so I got it fully installed. I mean, we might have to put Those. Those posts in the Truth Zone.

It looks pretty good on iPhone.

It looks good.

It's, like, pretty readable.

Give us the review.

I think it's good. I like it.

Okay.

I might keep it. I mean, it's a bit slow on my phone, and that might be because.

Which one you have?

I have an 11, so it's like.

Whoa, wait, I'm on, like, 16.

He's got a vintage iPhone.

So it's either that or I think, also usually in the beta versions, it's. It's a bit like, slower.

I have a 16 Pro. Yeah, I'm light years ahead of you. Oh, my God.

But okay, it looks pretty good.

It looks good up there.

Okay, cool.

Yeah, bring it up.

Bring it up. Let's see it. We're putting on the printer cam. We'll see it. And. Yeah, here we go. Okay, cool. Can we see?

I don't see much.

It actually looks pretty similar, I expect.

Okay, there's the. There's the. There. It's transition.

Yeah, this doesn't look as.

It's mostly just the home screen. Like, if you go. I went to a bunch of the, like, open, open.

Hits the calculator.

Looks. No, no, no, I'm kidding, I'm kidding.

You can see the glass effect, like, right here up at the top.

Yeah, it's nice.

And look. So look at the search bar as I scroll the apps here. You can see that glass effect. It's not too bad. You bring down the settings.

I'm bullish. This is groundbreaking. I think this is groundbreaking.

This works. This is pretty readable. Anyway, thank you for volunteering. You'll have to. We'll have to check in with you over the next couple days. We're sorry if anything works, if it.

Destroys your productivity, but you're cracked, so.

You're cracked.

You'll figure it out. Anyway, thank you, Tyler.

You know, Apple's going to be pushing this pretty hard. They got to do some out of home advertising. They got to get on ad quick.com out of home advertising made easy and measurable.

Drop the Genmoji ads market by every single billboard in San Francisco for iOS.

Tim, say goodbye to the headaches of out of home advertising. Only Ad Quick combines technology out of home expertise and data to enable efficient, seamless ad buying across the globe.

I'll be right back.

We have our next guest, Scott Belsky, coming to the studio in just a minute. In the meantime, we will react to some posts on the timeline. Kendall says, trying to quickly change the subject on the zoom call when everyone Else is cracking jokes about the Waymos. And I can feel my. I can feel myself choking up. Lots of sympathy for the Waymos. Lots of coverage in the Wall Street Journal today about the. About the burning Waymos, talking about how this played directly into the hands of the Trump administration by making the tech company clearly align with, you know, against the burning of Waymos. Very sad to see. Hopefully things resolve. It does feel like the report from la. I've been driving around today. It feels like things have calmed down a lot in general. I was able to drive all the way across LA without any. I was in downtown, didn't really notice anything and so it does seem like things are. Are calming down, thankfully. Anyway, I like this post from Tyler. I'm gonna mispronounce this name. Well, I'll just say Tyler, he says, unlike the rest of you cowards, I think the Y Combinator valuations are too low. We will find out what the valuations are coming in at tomorrow when we're at YC Demo Day live in San Francisco. We're bringing a huge swath of the crew. We're going up right after the show wraps. The team is stoked. We'll be giving out hats. We have a whole bunch of different hats, some TVPN hats, some ramp hats, and the ramp hats have TVPN on them. So they're all limited editions. I think people will really like them. So if you're a founder at YC Demo Day or you're an investor at YCDemo Day, stop by our table, come chat with us for a couple minutes and we will get your take and hear your pitch and hopefully turn your Demo day into a massive party round because we will be live from the palace of Party rounds. Mike Solana reposted the video of the Waymos leaving Los Angeles. Apparently Waymo is moving their cars out of la. I don't know how real this is, but it was a video of several Way Mos driving through la, kind of all together. Maybe they're leaving. Getting out of here quite makes sense because you think you could just put them wherever they store them and charge them and clean them.

But yeah, I don't necessarily believe that this was a response.

This seems like fake news. But Solana has a great post as always. The elves are leaving Middle Earth. Very funny. Anyway, Ramp has a post in 2012. In 2024, ramp placed 32 on CNBC's Disruptors 50 this year. Number six. They're grateful to their customers for betting on a better way and allowing us to Fix the parts of finance people hate so they can do more of what they love. And shout out to our friends at Anduril for making number one. What a fantastic lineup over there. What else is in the news? Oh, Warner Brothers. Oh, wait. We have Scott Belsky in the studio. Let's bring him in. How are you doing, Scott?

There he is. Finally. Welcome.

I'm sorry it took me so long to invite you on. I'm so glad we can make this happen. It's fantastic to find.

We were sending invites constantly.

Mentally, we were like, this is a dream guest. So thank you so much for taking the time to join us. Would you mind kind of setting the stage for us? Because you've done a lot in your career. What are you focused on right now? What's interesting to you? And then we can kind of dig into all the hot topics of the day.

First of all, it's been fun to watch you guys at work, and congrats on just, you know, creating conversation, which I think is appreciated by all of us. What am I up to these days? I mean, my obsession has always been the intersection of creativity and technology. Behance is probably, what, 58 million, 60 million creatives online right now focusing their work. That brought me into Adobe. Had a long stint at Adobe leading emerging products and design and various other new things there. And now I'm excited to kind of explore the world of storytelling and the role that technology plays there. Obsessed with the implications of new technology, you know, that emerge. And I love engaging in conversation about. You guys were talking about the implications, kind of writing I do and riffing off of, I don't know, like, what's going to happen? Because of what's going to happen is the question.

Yeah, it's a big question.

It's a big question.

Maybe we should start with some of the news that's coming out of, like, this idea of collective memory. What's going on in the enterprise we were talking about Glean OpenAI Deep Research. There's so many new products. We're using stuff, we're wiring up stuff. We have an intern in the corner vibe coding different stuff for us. We're essentially a podcast or a live show, a media company, and we're writing software. Everything about how we are building this media business is different than I think people would have done even just a few years ago.

But it's actually crazy. If somebody told me three years ago, we have a show, a podcast, and we're building a bunch of internal software, I would just be so bearish and yet we get so much value out of our internal software we call Newsmax Newsmax. But anyways, back to you.

No, I mean, I think the. Well, first of all, when you're building a new team from scratch these days, it's almost like you're able to build with another new generation of tools as opposed to retrofit something built with old tools.

Totally.

So I do think there's this advantage that a lot of new companies are realizing now, or new teams within companies. But you asked about this concept of collective memory, which I've been obsessed with, this notion that the context window of these AI tools that we're using, of course those context windows are growing, which means that their memory of us and their understanding of us is growing and persisting, existing across all the inquiries we have. And, you know, and this is becoming almost like, you know, an extension of our knowledge and opinions and identity and everything else. And it's great when these LLMs get to know us and give us better and better answers to our questions. What happens when in the enterprise, we access each other's memory? Right. So if the, if the context window of me working, you know, as one of your interns for years, you know, is suddenly accessible to you, even when I leave, you know, are you able to keep saying, hey, what does Scott think? Or, you know, or what. What. What did Scott do? And we had this situation, and is that a, Is that, does that belong to the enterprise? And you do see, by the way, like every company right now trying to build this memory for each of its products, users, and in some cases, a lot of companies are building connectors to get the data from other products that that customer uses to enrich that memory and understanding of their customer. And it'll be interesting when that kind of memory that's for each of us becomes a collective memory. And then what happens in the, in the social, like in the consumer world? Like, you know, does your girlfriend say, like, I want access to your context window, your memory, you know, and is that kind of weird? I think there's all kinds of weird questions and implications that arise from that.

Do you think memory can be the lock in that? I think certain companies would like it to be. It doesn't feel like it is yet, but I think there's still an idea that it could be.

It actually goes the opposite way. Like, I'll ask ChatGPT to tell me a joke and it'll be like hyper specific about something I talked about a year ago, and I'm like, that actually makes the joke worse. ChatGPT. Like, I'd prefer if you ignored all my questions about trains or something.

Well, that could be in the prompt, right?

Yeah, exactly.

I feel like, you know, you see OpenAI launch the ability to log into tools with OpenAI and you know, when I saw that I was like instantly thinking, gosh, that's a way to enrich the personal personalization. I, I do think personalization effects are the new network effects, you know, we used to talk about in terms of the, the moat that companies would have with a network effect. Today I do think they will have with personalization. And that is a direct, you know, a direct out outcome of the memory and extended context window and all of the data, you know, that they can get to give us a personalized experience that no one else could ever give us. But you're right, like there are other ramifications of that if you're too well known. I asked OpenAI the other day what I didn't know about myself and it told me that you're a real hypochondriac. And I was like, well, I didn't know that about myself.

That's funny. I want to know more about this idea of collective in the enterprise.

Sorry to interrupt, but it's an interesting thing where the model is effectively trying to do pattern recognition and it just identifies something that is.

You ask a lot of questions. It's like, yeah, because you're a question answering machine. Yeah, like that's not me, that's you. But on the concept of like creating a collective memory in the enterprise, I've noticed that even though we are this new company that's starting from scratch, total greenfield, we have still not been able to go all in on one walled garden. We have Gmail, we send a lot of imessage groups. We also have a slack that, you know, we are interfacing with different people on. That's a Salesforce product. And so I'm wondering if, you know, there's been Eric Mykovsky was trying to build beeper this product to let you create one unified messaging interface. That was a real struggle because every major company was like, we will not let you break our walled garden. We're not really there yet. And it seems like most of the companies are playing ball with MCP servers and the like. But what do you think the big dynamics of the big tech companies will be like once they realize that, hey, maybe if I don't want to let the fox in the henhouse.

Well, I think you're, you know, you're, you're you're forecasting a bit of a, call it a data war, you know, in the years ahead where everyone's going to leverage each other's APIs and start to pull all the data, you know, in. And they're going to probably be some backroom board level bilateral negotiations where it's like, well, let's not cut off slack because we don't want them to cut off us. So we're going to let them. And may the, you know, may the best, may the best model win. Right, with the data provided to it. So, but that means that there, if the data becomes sort of readily available across the board from all of the connectors that are made for all of the products we use. And to your point, if you've built the network, your, your company with all these different products, that's fine because the data is all stored probably in one of three or four clouds and you can just instantly access all of that data, use AI to see it as structured data, and then be able to have AI apps built on top of it. And you don't even need to care what actual products create the data as long as you have access to all the data. But then the question is, what are the moats, right? I think one of them is the personalization layer that we talked about just now and that context window becoming ever more important in the product. I think another one is permissioning. In the enterprise especially, permissioning is fricking hard to know exactly who's allowed to ask what question and get what answer or what column of information that might relay itself to that answer. That is a very, very hard thing to crack. And I do believe that the enterprise products that have really incredible permissioning layers and products that allow people to control and access permissioning, you know, will also also have an advantage. So I mean, and then in the operating systems, we talk about operating systems in the consumer world with, you know, your iOS or Android person or whatever, and that's the operating system. Those, those operating systems, like if they can't figure it out, well, shame on them. Like they literally own the top layer of everything underneath and they should be able to manage our attention and our needs with AI at that level. But what are the operating systems of work? I think there's going to be another war played around those operating systems. One might be the browser, I think one might be the way that projects are managed. Different functions of the enterprise might have different operating systems and people are going to be fighting to own the top layer of those.

Yeah, I wonder if that's like a, almost a bull case for data lakes becoming more of a mid market or early stage product. Because like I take your point about like the deals between the large tech companies just kind of happening behind the scenes. But at the same time I'm just like there's going to be friction and I just know these companies like yes, I can export all my data from Facebook if I click 12 buttons and do it every single day. Like the API is not fully there for a lot of these products. And so I wonder if more and more people are going to be saying like even for my personal life I want to be able to funnel everything into a data lake that then I can drop my own AI on. And I don't necessarily want lock into one AI tool. Right now I'm wondering how that, how that will evolve. I don't know.

Yeah, I mean we're going to see this in consumer too. It'd be really nice right now if you could describe in ChatGPT, hey, go find this image and then make it a studio Ghibli. Right. Or make it like an impressionist painting.

Right? Yeah.

And obviously there's that kind of barrier.

Yeah. It just feels like the big tech companies want, they're fine having lock in and like you see this with wwdc, like the Siri button cannot be remapped still. And it's like that would have been an easy thing to do. That would have been something that a lot of companies, consumers demanded. What was your reaction to WWDC and kind of like how, how the, the AI Apple does, like decisions are playing, are playing out.

Yeah, I mean I think it's really, it's a really fascinating time, you know, in that company's history. I think from the outside we're questioning like where's the vision, you know, where are we all going? And I think inside it's probably some sense of let's wait until we're really ready. We've had a few false starts. We can't get it wrong again. And so we're probably going to get something that's more fully baked which is hopefully great for Apple. You know, I never, I never discount Apple because of its DNA and its rigor and the quality of talent that's there. You know, on the other hand, like they're late to the game at this point, like super late. And I also wonder the things we discussed just a minute ago around personalization and some of the moats that are, they're going to play also by the way, remember this is a new era of experiences that is all about data. And Apple's always been all about privacy. So Apple in some ways as a policy has done everything to make it impossible for others to get others data. And here we are in a world where that's what's going to enrich the experience we have from AI tools. So I think it's a, it's a different Apple.

Do you think that Apple will end up looking really smart on privacy? Like we're in a very chaotic time where people will let a new software, you know, company just have full access to recording visual screen, recording, record my screen.

That's like it does okay thing now.

And I'm even surprised at this point that we haven't. I can't remember a major incident of a model producing private data in response to, to sort of like a general query.

I've heard about anecdotes here and there, but nothing really.

You can imagine a world some days where, you know, people get their kind of totally lines crossed and you ask for, you ask about some, some, you know, acquisition, for example, and somebody at some point dumped you.

Even tried to get chatgpt to guess a picture that was from your own house. And it was, and it was smart enough to say, hey, you're maybe trying to spy on someone. So I'm not going to do that.

Even though it would basically gave me like the exact description of it without giving me effectively the street.

So it knew, it knew not to step up the line. I don't know.

Anyway, on privacy, it's a great question. You know, I always go back to first principles on this one and I say to myself, you know, first of all, what do we know about the next generation? They seem to care less about privacy than the generation before them.

Yeah.

You know, number two is we're willing to trade a lot for less friction in our lives. And it seems like people are willing to always click accept or yes or yes or yes again just to be able to have a frictionless experience in a product. Now I think you're surfacing the point that we don't oftentimes realize how much we're giving over in those instances. And I think the world of AI will make that very apparent to us because will start getting these personalized experiences. We're like, how the hell do they know that about me? Right. But I don't know, like, there's also a part of me that says that the best technology takes us back to the way things once were, but with more scale and efficiency. Well, the way things once were hundreds of years ago is we were known like everywhere we went in our small towns we were remembered. They called us by name, they knew our kids names, they knew our favorite cut at the Butcher. Like you were kind of known. And in some weird way we're being brought back to a world where we're going to be known again. The question is we want to know how we're known and also we don't want to be known by people that we don't trust.

Yeah, yeah. There is kind of this interesting economic dynamic where there's almost like a de facto bug bounty for privacy in the fact that if a company like Apple has a privacy fiasco that could wipe off hundreds of billions of dollars of market cap and so they have this massive incentive to invest to not have these privacy fiascos or the data to leak. And so you have kind of this massive economic capital cannon towards let's secure the public clouds. And so one of my former colleagues was making the point that in terms of open source AI versus closed source AI, oftentimes it's actually better to trust a hyperscaler with your data that because it's harder to break into an AWS data center than it is to the on prem server that yes you're running it but you're the only security guard. And what happens if someone breaks in and just takes the rack?

I'll tell you what though, like here's a bullish case for Apple would be that models are all going to be local, you know, in X number of years. The most power, like if you look at the performance of LLMs, right there are these small LMS and there's these ones that run locally on devices. And the, the, the, the, the delta between those and top LMS is of course decreasing. And so as chips get better and LMS run locally, you can imagine that all of your data can actually be stored locally in a super encrypted way. All these local models can operate to answer all of the questions you have, can search the Internet when it, when is necessary and maybe we all opt to this like super private AI world maybe for all we need Apple's optimizing for that flag that they know the Puck is going to, you know, and they're going to launch who knows how many years from now. Like 100% local AI experience that is world class with zero privacy considerations or risks rather.

Yeah, yeah, yeah, yeah. That makes no sense. Can you talk a little bit about knowledge arbitrage? Like these models we've gone from you know, massive Databases of facts. And so there wasn't as much incentive to just be the person that memorizes all the facts because you can just Google it on your phone. Now we're in this age of even more significant knowledge retrieval intelligence, potentially to chip to meter, if we can call it intelligence. What is knowledge arbitrage?

Yeah, I mean, it's funny, when I think about the term knowledge arbitrage, I'm actually brought back to. I think this is very relevant for people who are listening to us or kind of earlier in their careers right now and figuring out how to plot their path forward. You know, you go back to the early 2000s, and the people who stood Twitter and social media and things like that were getting hired by like, CMOs of Fortune 500 companies to school them on what the heck this thing was and like how to have their brand participate. And it was really looked upon as like this complex, crazy thing that every leader in corporate America felt they needed to understand. And so there's this moment of knowledge arbitrage where people who are like, you know, in their late teens, early 20s, who are deeply, natively familiar with this stuff called social media, could just like school the people who had no clue. Here we are again. You know, you have the leaders of all these Fortune 500 companies saying, oh my gosh, like, we have to refactor how we work and every function is going to operate differently with AI, but we don't even know, like, when to use chat GPT. We don't even know what a prompt is. And like, how does this stuff even work? And so once again, these like, very young people who got through college using ChatGPT and talked to ChatGPT about their boyfriend or girlfriend issues or whatever else, like it's native to them. And so I do think this knowledge arbitrage moment, you know, is here. And the window's open right now.

Yeah, it's funny that there's this massive fear from, from young people around how the job market is evolving, are jobs getting killed, whatever. And at the same time, yeah, it feels like social media didn't reinvent every part of a company. Like, it sort of massively changed marketing and how you should focus marketing dollars and how you should think about building brands and all these things, but it wasn't something that changed internal comms, like you're talking about with like collective memory. It wasn't something that changed a certain aspect of the business, like compliance. It didn't change, you know, it wasn't this sort of like full stack transformation. And for that reason, it creates more opportunity than the social media wave to just get good at using alarms and agents and whatever ever, whatever other tools you have, and then use that. I mean, we. The last person that we hired was somebody who vibe coded like a guest directory for us. And they did it last, like Friday, last Thursday.

And then we hired them.

Hired them that. And then we saw it, like, Friday. They started on.

They came on. Yeah, they came and it was on Thursday. They came on Friday and we had them.

And just by demonstrating the ability to leverage these tools and create a product that was a V1, but it was a solid V1, and it was just a super easy decision to be like, yeah, you should join the team. So shout out, shout out to Adam.

But like 48 hours work, though, that's like, incredible. Yeah, so by the way, that's changing the way product is even done. Like, I'm so used to a world where you concept, you do like a sort of, you know, product scoping session, then you have the designers work in a product like Figma and make prototypes, and then engineers don't even see it until the prototypes are locked. And then it's like redlined and engineering start to, like, you know, now it's a different world. Right?

Yep.

It's like, we need, we need, we need this to be done. You kind of vibe code it out in parallel. A designer says, okay, let's how to finesse it. Here's how we can maybe change that, make that more accessible to more people. And then you just deploy and test and iterate. It's a different, It's a whole different product stack to some degree. And that's why I always say the best talent these days is like a collapsed stack talent strategy where you hire people to have exposure to different parts of the stack, whether it's engineering and design or design and copy or product and design. Because you're just trying to collapse the stack with the types of people you have so you can leverage these two tools, you know, efficiently.

Can you talk a little bit about agency? It's, it's an interesting concept because we don't have like an eval for it in terms of the LLMs. We're, we're, we're, you know, destroying all the benchmarks in math and iq. But the models don't seem to have volition. And yet the trend is around agentic applications. We're trying to give them agency. It still seems like there's some sort of combination of uniquely human traits that seems to be increasingly valuable. How are you defining it right now?

Well, I think that when you talk to, when you talk about sort of agentic workflows and the role of agents in the future, and I think about it, every company I'm calling companies these days, in my mind, cognicos like cognition driven businesses where it's ultimately inference in the center, right? And they're these, the functions of a company which were once accounting and legal and design. And like all these different functions are going to be AI tools that are running, you know, AI agents basically that are running that function, or a collection of agents doing different aspects of the function. Which then begs the question of what's the role of the humans? I think we are the orchestration designers and orchestration engineers. We're like orchestrating how these agents work and how they work together. And we're also developing, enforcing the rules by which they work because remember, these agents are told what to do and they're rewarded based on doing it. They have to be working within a subset of rules that, that need to be administered and enforced to some degree by humans. So I, you know, so I start there and then I start to think, okay, so what are the functions of every organization going to look like? You know, what. And then that dictates like what kinds of tools or agents are empowered to, you know, quote unquote, take agency in those functions. But I also want to say that when you say agency at first I think of like the human, you know, the human getting.

Still got it.

We gotta, dude, we gotta, we gotta exercise our agency and our taste more than ever before these days and the years, the days of us kind of getting a prescribed job function and sticking to it. That is like the deathbell of a career these days.

Yeah, totally. As soon as it can be defined, it can be reinforcement learned again.

Were you surprised at all about the concept of taste coming into the discourse, the Silicon Valley discourse? It feels like one of those things. It actually reminded me of when venture capitalists discovered the creator economy in 2020 and 2019. But it was a trend that had been happening for like a decade at that point where people were creating content online and turning it into, into a business. And then all of a sudden people I think saw a statistic that was like, you know, look at this, you know, look how fast this category is growing. We should invest a bunch of dollars there. But in many ways I feel like taste has always been a key part of our industry. And if you look at a lot of the best companies, they don't all look the same. But they have people that run them that have taste, that that sort of gets applied in many different ways, whether it's hiring all the way through. Obviously the visual side.

You know, I am brought back to a. Just earlier in my career, you know, it was all about what skills you had, right? And the resume would always, were you good at PHP or JavaScript or could you do wrote code in Python or whatever the case might be. We come from a world where skills like, were the hardest things to achieve and the most differentiating things for humans, right. In a project now, as a lot of the skills have been offloaded to compute, you know, now it's like, well, what's left? Like, taste is actually one of those things that is definitely left. Right. And so I think that the fact that everyone's sort of focused on taste these days is more of a commentary on how less focused they are on skill and how much confidence they have in all these models and coding tools and everything else.

Can taste be taught, though? That's the question. Skills can be taught.

Like, let's, you know, can taste be taught is like a thing I think about in debate all the time with, with friends. You know, what is taste? Gosh, it's like the human experience. It's childhood traumas, it's mistake of the eyes, it's judgment, it's knowing how many things to see before you make a decision, you know, but it's also knowing how to choose something that isn't so on the nose but like leave something for the imagination, you know, it's knowing where people are going as opposed to where people are. Like, I think it's all of these factors that are so critical whether you're making a product or making a marketing campaign or a film or whatever the case might be. And I just, you know, I think that hopefully we'll just over index on human experience in order to achieve taste.

Yeah, there's. Yeah, it almost feels like just independent thinking. Confidence, like you can maybe develop.

No, but it's independent, Independent thinking that elicits a collective response that you want in some ways. Right. I mean, I think of this in the context of how companies get attention. There's a lot of different ways to get attention. Some, some, some ways are, you know, some ways are tasteful and elicit. I know what you're thinking of, you know, positive response. Other others get, you know, a lot of attention. But, but a sort of visceral negative reaction. And taste in the context of social media algorithms and timelines is like such an important, important question because it's really easy right now to do, to instantiate.

Something, but not to. Yeah, it's fascinating.

Yeah. Taste, Taste too in the context of AI Slop. Right. There's tasteful, there's tasteful.

I always keep coming back to the Harry Potter Balenciaga video. I feel like that was, you know, it, it had the aesthetics of AI Slop because it was, AI generated issues with it and yes, but it, it had the, it had the touch of a human to think of combining those two kind of completely unrelated disparate ideas. And when you put them together, it was something that you had to watch.

And by the way, that's perfect evidence the fact that you're still citing that despite how many months, if not years of content generation has happened since. Yeah, it shows you like the taste still stands out and that's why, you know, people are like, oh my gosh, content is commoditized. Like everyone can do everything now. Oh my gosh, this is horrible for Hollywood. My view is the opposite. Like in the, in historically when anything becomes commoditized, whether it's shoes, clothes, handbags, liquor, whatever case, like that's lvmh. Like LVMH is in all those businesses. But they're doing the higher end, meaning infused, more scarce version of that thing that's available to anyone. Yeah, your now content is commoditized given all these new tools and I think as consumers of content, we are going to seek more scarce, meaning infused, human, you know, brilliant brand signaled versions of content to fill our attention in time. And so I think that's great example. You know, it doesn't matter how it was made, it matters, you know, whether there's meaning in it.

I even knew how the tools, I knew the tools that were used to make the Harry Potter Balenciaga thing, but I couldn't come up with a unique idea that would go viral. And so I, I went to ChatGPT and I said, here's why this worked. It worked because it was inspired and it combined two things that were so disparate. Think of another thing that could be iconic like that. And it couldn't do it at all. It came back.

It was famous.

Yeah, yeah, exactly, exactly. It wasn't there. You know, that inspiration still needs to strike.

Jordan, do you have your last question? Speaking of Hollywood, did you see Mountainhead, the new Jesse Armstrong film? Have you seen it yet?

I haven't seen Mountain Head yet, no.

Okay. You should. I mean you, if you don't enjoy it, they're really cooked. But it's Jesse Armstrong from Succession. You're making this film. The high level concept is, you know.

Some be all in podcast.

Yeah, yeah, you could, you could say that. But the thing that, the thing that I wanted to get to is basically the overarching part of the overarching narrative is that deepfakes have gotten so good that they're causing global chaos because people don't know what's real. And then. And they're seeing video and sort of acting on that video. I wanted to ask you in that context or just generally about the progression of video models and how you're thinking about, you know, if VO3, what do you expect from things like VO4 and you know, new Runway models and any, you know, I'm assuming you're looking at everything internally.

Yeah, let me just. Well, two comments like one on video models and then deepfakes real quickly on video models. I mean this is like a consistent slap a hand game, right, where oh my God, this is the best model. And then VO2 comes out and then Sora and then Runway and the VO3. And it's just, it just keeps piling up. This is great for anyone who makes stories because these models are becoming better and increasingly commoditized. However, you know, it's. Would I want to be the model, you know, that's competing in that increasingly commoditized and not as great, you know, quality sort of vector actor. Probably not. So that's comment number one on the deep, on the deep fake situation. It was interesting. Like these new, these new websites that are coming out where, where you have these, like anyone can make a deep fake with Trump saying whatever, you know, I actually think they're playing a secret value to society by inoculating all of us.

Totally.

We all get a fake Trump said something crazy video tomorrow. And we're inoculated from the fact that we should no longer trust what we see. That we should, instead of going from a world where we would trust but verify, go from a world, you know, get to a world where we verify, then trust like that would be good for humanity. So I'm like, I'm actually, you know, for that inoculation because I think, I think that, I think that we need it and we're going to want to know where, where stuff came in a weird, weird way.

It didn't happen.

No, in a weird, in a weird way it might save legacy media because if legacy media oriented institutions, it orients around, like if CNN were to say we are, you know, there's a bunch of These, you know, you have usernames like Autism Capital that are sort of de facto news on X. Right. These are accounts that get billions of impressions and they act as, as, you know, news content. But if cnbc, you know, CNBC and CNN and other services can really orient around. We did the, you know, we were late to the punch on this, but we did the work to verify that it was real. It actually might save totally.

And there's credentials now, there's metadata through things like content credentials that go into assets that are made on certain cameras and edited in certain software. And you know, and whether it's TPPN or CNN or anyone else can surface that information in the reporting and say we verify this is, you know, done by the so and so and was edited by so and so. And I think that adds to a layer of what people are going to need and want and future. So thank you so much.

Yeah, I wish we had a. I Wish we had 90 minutes. Yeah, this is super fun.

We have to come back soon. Yeah. To be continued.

Great.

Have a great rest of your day. We'll talk to you. Have a good one. Bye. Cheers. Really quickly, let me tell you about bezel. Go to getbezel.com, your bezel concierge is available now to source you any watch on the planet. Seriously, any watch. And we have our next guest already in the studio, Roy from Bloomberg Beta coming in to talk about his latest fundraise. How are you doing, Roy? Great to meet you.

Fresh capital in the bank or committed or something like that.

How much, how much capital do you raise?

Talk to us.

Okay.

How much?

You hit it 75 times for 75. You know, I'll take one customer sale over Fresh capital in the bank. You know.

Capital formation is the highest calling of man.

It is they two guys who do.

Not intend to form their own capital unless.

That's right. But we, we're in the business of covering formation, celebrating it.

But congratulations. What are you most excited about? Where are you planning to deploy this? What is, how is the money burning a hole in your pocket?

Yeah. So we 12 years ago started this firm to say we're going to do day zero investing in startups. And I'll just give the context because the answer is going to be the same thing. Startups that future are future of work oriented. Our bet at the time was our personal lives have changed a lot. Our work lives have not. Maybe that'll change. Like people were still using Lotus Notes for email and. And now on our fifth fund, we are doing Basically exactly the same thing. The craft of Day Zero. Work closely with the founder. And what's burning a hole in my pocket is continuing to do that, you know, And I'm looking at our Fund 1 companies and feeling like they're still midway through the journey. Replit Flexport Campus. You know.

We love these companies. Most of them have been on the show. Fantastic companies.

Yeah, no, I've loved watching them. That's part of what I love about the show.

What percentage of funds, venture funds, make it to fund 5?

That's an excellent question.

It's gotta be like 5%, very low.

But I always feel like that's sort of like those stats about what percentage of startups, like raise their Series B. It's like if the starting denominator has a lot of meh in it, like, I think to me, the more interesting question is who gets. There you go. Who gets to have. There you go. Hot off the press.

I printed. So I printed your. Your post because it appears that you printed your post. I don't know if you can see this, but you're going meta.

It's a print.

LinkedIn, I think a LinkedIn post. And you marked it up. You, you said, today we're announcing our next fund, our fourth, and you just crossed it out and put fifth. And we respect printing out posts here. So I just wanted to.

The idea was inspired by this thing I stole from Congress when the Republicans blocked a Democrat Supreme Court nominee. And then it went around the other way and they literally just crossed out the name, put in like Schumer to McConnell. It is the same. So I think the more interesting question is who gets to a Fund 5 or whatever fund size without trying to gradually expand AUM? Like, it seems like so many of the VC firms out there, and we love working with them, are just in the AUM expansion business.

Yep.

The big AUM scare me personally. I looked it up here. According to Chad GPT, there's probably a 50% chance. It's hallucinating, but it's estimating at less than 5% based on PitchBook and Cambridge Associates data.

I mean, here you say you aspire to be the most transparent investors. What does transparency mean in the context of a venture fund?

Yeah. Do you tell the founder, look, we clocked you at about a 15% chance of success. It's a bet we're willing to take.

You try to tell them how exactly we're thinking about it. But here, look up, go to our website right now. Those two laptops you got in front of you and then you'll tell me what. But. Which is to say when I was a founder, so before this I started a company in games and you know, you drive down to the 280, end up in an office on Sand Hill Road and in the first three minutes of the meeting the person would be like, oh, but you're not a fit for us. We don't do this. And it's like, well, why on earth do we just do that? And we're an attacker. We're coming into an established industry. We just asked, are there things the industry does that we can flip the bit and do the other way? And so the industry seemed very opaque at the time. It's a lot less true now. And we're just like, can we be as transparent as possible? Can we put in valuation? Can we put in diligence questions? You can actually get our long form actual deal documents from our website. And the point of it is not transparency for transparency's sake. Although that's a very Bloomberg idea about transparency. It's just founders are our customers. We want them to not waste time guessing what we do. And so if they can disqualify out by seeing we don't do that thing, great. Better for everybody.

We're big fans of Bloomberg. We read a lot of Bloomberg pieces on the show. What is the actual relationship?

Do you have your terminal subscription yet?

No, we got to get one. This is my go to Terminal. Every day is a Wall Street Street Journal. But we, we, we do have a.

Description and you gotta get a term.

And we've had a bunch of Bloomberg folks in the show. You're like the fifth Bloomberg affiliated person. We got Joe Wiesenthal, Tracy Alloway, we got Shirin Ghaffari coming on the show.

Bigger brands than me one and all.

But what is the relationship between the fund and Bloomberg as a whole? Obviously Bloomberg does a ton of different stuff. What does that look like and how is it involved?

The company is the LP in our end of story, full stop. We are not a strategic investor. So unlike a bunch of other corporate backed funds, we're not trying to find some startup that's going to be a Bloomberg partner.

Sure.

Bloomberg wanted to understand what was happening in the world of startups. The argument was best way to do that is give startups what they need, which is just money.

Yeah, that makes a lot of sense. Cool. Talk to me about deployment of the fund is 75 million. That's a lot of smaller checks. Are you holding back some for pro rata and scale up like you know.

The line about like tell me the fund size, I'll tell you the strategy.

Yeah.

Part of the reason why we've kept our fund size where it is is we want to be as early as possible. And so I can't remember the exact number, but it's something like most funds, their first check ends up being something like 75 basis points of fund size on average. And we end up a little bit around there, slightly bigger. We do, you know, we reserve for follow ons. But then we also have an opportunity fund because what we realized is we wrote some big checks from our core fund into some of our winners. We wrote a big check. Yeah, interplit at one point, whatever, you know, company new front. And it's like, well if we're going to do that, we might as well have the capital to do that reliably. So we've got an opportunity fund.

Yeah. At the early stage. Are you, are you trying to pick a winner and then not invest in competitive companies? Like how do you think about that?

Very old school and all this stuff. My view on competition is in the eye of the founder. Like I have an attitude on this which is if I back you, I want you knowing I have your back, not that I'm thinking about which of your competitors to introduce to somebody. And so any founder we invest in, if they tell us that a company is competitive, we won't invest in that company. Shit still happens. Like of course, you know, companies competitive.

And we've got to work it out.

But that's the principle.

Yeah, yeah, but, yeah, but that's not on you driving the news cycle. This week you got wwdc. Apple kind of pulling back the news cycle.

It drove the news cycle kind of by necessity, but kind of. If you had said but nothing happened, would we have just nodded along politely?

Well, our intern Tyler spent hours installing it.

Well, yeah, he assembled the first American made iPhone yesterday. Today he installed arguably a harder challenge installing iOS 26. But I mean there is a bit of a narrative for startups there in the sense that if Apple is truly pulling back and not going to steamroll the startup developer community in terms of AI based apps, that could be a potential boom for, you know, we could see another rise of a mobile esque era where AI companies are building on iOS in a little bit more friendly territory than it's been in the past. Obviously there's the Fortnite decision and discussion around the App Store. What are you seeing that's interesting or just what else is driving the news cycle? For you? Well, no.

So I do think, look, Apple has to be an attacker on AI. You've pointed out the way that the privacy first philosophy makes it a lot harder for them to do AI. And AI developers who go there talk about running into that internally. You know, all the big tech companies want to be the platform on which everybody else share crops. And at the moment that's OpenAI and anthropic. It's like you have all these fast revenue growth.

Sharecrops is such a savage but real.

Way to describe a 30% cut. It's just 30%.

Just give me 30.

Thirty, of course, naturally. But the market didn't exist. But for that, yes. And, and I think they're going to try to figure out can they disrupt some other part of the ecosystem. I feel the same way about the scale AI acquisition, which is meta has to attack. They need a layer that they can control that everybody else depends on and this looks like their play at it. The other thing I think is interesting about that, that I've been thinking about is how many of these big AI companies have effectively been acquired in ways that the government doesn't get to approve. It's, you know, the big talent acquisitions. When I saw this, I was like, that looks like an acquisition the FTC doesn't get to approve. And so I don't know if that's the motivation. Have not spoken in any of the parties. We're not in scale AI, but my sense is you're basically seeing the Game of Thrones play out, but the government is one of the houses.

What's very interesting about that is that we are, yeah, Lina Khan's gone and we're in a different regime, but it's the same thing.

Let us, let us do some acquisitions. Let us, let us do some, some M and a.

Just totally 28 billion dollar acquisition. Just let us sneak it, sneak it in. It's fine.

Let's get it by. Let's get it just one. Let's open. Let's get the M and A window open.

I mean, the other thing I say is, when was the last time you heard a big tech company buy a company for $80 million? It felt like that used to happen all the time.

All the time. Yep.

And now it's a lot less common for that.

They took that from us.

They took that from us.

I think I feel the grief.

Yeah. I mean there used to be, there used to be a really great outcome for a lot of folks where if you're a founder and you start and you still own 20, 30% of a company and you sell it for 80 million. Like that's, that's a house in Palo Alto and kids and generational wealth in some ways. And yeah, that has gone by the wayside. I don't exactly know why because those seem to be easier to get across the finish line from an FTPC perspective. But I guess folks just.

Yeah. I mean at what point does the FTC just say like, look, we know we could see what you're doing.

Yeah.

I mean also 15 billion on this company for. And the CEO no longer works there.

Yeah. I mean the flip side. Yeah. I mean the flip side is like when was last time a company was valued at 80 million? Because every company seems to go. It was straight from 5 million to 500 million in two days. And so I don't know where you're getting 80 million. That's just hard to. Hard to find. I do, I do have a question for you. I don't even know if you can speak to this, but there's a lot of companies that are thinking about plugging into the terminal. We see what Perplexity is doing. We see a bunch of AI agents.

For Perplexity is trying to destroy, disrupt the terminal.

Are you looking at any of that space? Would that be competitive? Is there anything interesting there that you've seen or heard rumblings on the other side of the fence?

I can't speak to what's competitive with, you know, they're my lp.

Yeah.

Of course.

For their business strategy. What I will say is when you, when you run a firm that's got a name of another company on it, my guess is the Google Ventures GV guys have this too. Every pitch that's like we are going to be the Bloomberg for X lands on my desk.

Yep.

And maybe not everyone, but many of them do because they're like. And how amazing would it be to also have Bloomberg as an investor?

Yeah.

Almost all of them seem to misunderstand to me what the value of the Bloomberg terminal is.

Sure.

Because it packages data.

Yep.

Community with a messaging system, outside in perspective news, etc. I mean that's just some of it.

So much stuff.

And it's more than just let's display some data. So everybody who's trying to disrupt I think is in for a surprise when they realize what an ecosystem it is.

Yep.

I think this happens constantly, especially in media. People see a media product, they like it, they don't understand why it's successful, they try to reverse engineer it. But if you don't Understand the why you're never going to be.

But can I ask you guys though, for you, you tried before, what's your hypothesis on your why of why this has been so successful for you?

The main one is that most people in tech media have done it part time historically and we decided to just kind of burn all the ships and make this the full time thing. We got some great advice from David Senra, the host of the founders podcast, another full time media creator. And he basically just said, you guys have something that's working, you should go 10 times harder at this.

And what's interesting, and we also enjoyed it a lot.

We loved it.

Talking about business.

Yeah, we loved it.

Which was kind of the foundation. But yeah, the idea that there's a lot of, there's a lot of technology podcasts, there's a lot of business podcasts. If you actually drill down and you look at how many people are actually full time just making media, it's very, very small.

It's pretty small.

And the nature, the other thing is the nature of. By taking something way more serious than anyone else, it can sort of morph and evolve in the way that the show has from us to. In a printer.

My viewers look on this is. It's. It felt. I'm obsessed with Hamilton for a variety of reasons. One of the reasons I'm obsessed with it is nobody would have said one of the great media properties of the last 10 or 20 years would be a musical.

Yeah.

So it defied the genre through intensity and seriousness. And you feel immersed in a world and like before the show they used to go out on the street and do some little ditty or something like that. Oops, my camera just froze.

Okay, there we go.

And you guys have that same kind of immersion. I teach this class at Berkeley on the structure of the media industry. I will be inviting you as guest speakers.

Amazing.

We'd love to do it.

Not when you're live.

I grew up, I grew up in Berkeley.

The other thing that you've realized is it's cheap to start a media company. So as long as you don't get too big for your britches and figure like a venture backed company, your financial aspiration, maybe this is yours, is to create one of the most financially successful businesses in the history of capitalism, which is in VC. As much as an $80 million exit along the way can be magical for, for everybody involved. That's the hope is, you know, is that you can produce something wildly valuable and make some money.

Yeah, yeah, exactly. Yeah.

This is super Fun.

This is great. We'll have to have you back.

Congratulations.

Appreciate what you're doing. Keep us all immersed in it.

We will. We will come back soon. Yeah, we'll hit it.

We'll. We'll wait till after 73 more to go.

73 hits.

I'll come in and bang on.

Have a good one.

You're the man.

We'll talk to you later.

Cheers, Roy.

Bye. Up next, we have the founder of Metropolis, Alex Israel. He's in the studio. But really quickly, let me tell you about wander. Find your happy place. Book a wander with inspiring views, hotel, great amenities, dreamy beds, top tier cleaning and 247 concierge service. It's a vacation home, but better. Welcome to the stream. Alex, how are you doing today?

Good.

How are you guys doing?

We're doing great. We've been having an awesome show.

What's it like to land on the. The CNBC list? Yeah, congrats. I mean, that's got to be your guys's biggest accomplishment ever. There's nothing. There's nothing. There's nothing that tops that. No, no, no, I'm kidding. It's. I mean, anytime, you know, I know the CEOs that land on these lists, they're like cool. You know, they'll post about it, but it's like, you get back to work.

Yeah, of course.

Yeah, it's. Look, it's flattering. I just like being in the company of, you know, OpenAI. Yeah, you just got to get back to work.

Yeah, the first, the first CEO to be like, you know, normally CEO say job's not finished. In this case, job's finished.

No, we're done.

We're done. We're just done. We're done. This was the whole goal the whole time. And now we can retire. It's kind of like an exit, but better. Yeah, it's definitely better. It definitely, Yeah, I can pay my college.

It doesn't pay the bills, but you know, you can send it to a family member and maybe they'll be impressed, you know.

Yeah, yeah, you'll get some text messages. Anyway, can you kick us off with an introduction on, on, on the structure of the company and what you guys are building? Because it's super interesting and I know we can go a bunch of different directions, but I'd love to get your kind of ground setting first.

Yeah, of course. So, you know, traditional in a lot of ways. Startup Story founded the company in 2017 and now are one of the fast growing payments companies in the United States. We talk about ourselves externally as an AI company for the real world. But it's all about how you leverage computer vision and artificial intelligence to create next generation payments and commerce everywhere you go. Whether it's a retail environment, a car wash, or a gas station. How do you just walk in or walk out or in this case, drive in and drive out.

That is fascinating. Talk to me about the strategy of actual building the business and acquiring assets and actually getting like the go to market motion is different here, right?

Yeah, it was very different. I mean, look, we started very, you know, I don't know, stereotypical. We raised the, you know, I Guess Large, but $20 million seed financing, $40 million Series A.

We hit the $20 million seed in 2017. That, that, that, that's more than bigger than a mango seed, which, which started to become popular a bit later. I don't know. Is that a water, Watermelon seed round? What, what, what, what was the catalyst there? You just, you had a bunch of space and so people had a lot of confidence in you. Why did you need 20 million at the time?

Yeah, it's a good question. I don't know. I guess serial tech entrepreneur.

I don't know.

My last series A, in my Last company was 3.5 million. And I, you know, I thought I was rich, you know, and then, you know, $20 million seed financing. But look, we knew we were tackling, you know, the first vertical that we entered was mobility. And we wanted to go after the parking industry and we wanted to deploy technology in the parking industry. And we knew that there were a number of companies, whether it was Standard Cognition or Amazon Go, that were leveraging computer vision to create these next generation experiences. And we knew it would take a lot of capital and we knew it would take a while to penetrate the market. So we wanted to make sure that we had the capital in place to really go after the market.

Talk about your position in parking today. It sounds like you're thinking about other categories already. What point did you decide that you could be thinking about other applications of the tech and infrastructure?

Yeah, it's a good question. You know, I'd say so after our series A, I would say we hit product market fit and unit economic fit really quickly. I'd say the problem was we hit a wall with our go to market and that was, you know, at the time, we were scaling our technology to a number of the largest asset owners across the United States and we'd effectively knock on their door, ask to take the keys for the parking experience to their $200 million development and they were like cute startup, come back in 50 years. So we needed a different go to market motion. So we shifted strategies and we invented what we qualified internally as a gbo, which was this idea of a growth buyout. Could we raise venture capital dollars and could we acquire world businesses? So we started rolling up old world parking operators that are, you know, the cost plus staffing agencies that were even a positive that run parking across the United States. And we kept rolling them up to a point where now we're the largest parking operator in the United States.

That's amazing. What? Yeah, yeah. What is the structure of the, of the parking market generally? I feel like I live in Pasadena when I drive around. Some of them are city owned, some of them validated, some of them don't. It feels very fragmented. I couldn't name one unified brand. But is the structure of the parking owner market as fragmented as I perceive it to be?

Yeah, completely.

It's.

I mean, it's as fragmented as the real estate owner market.

Right.

People that own real estate own parking. And there are, you know, hundreds of thousands of real estate owners from a class A perspective all the way through an airport across the United States. And we deploy our technology to facilitate that type of seamless experience where a consumer can just drive in a text message when they arrive and get charged when they leave.

Yeah.

Is that historically just because there's no economy of scale like there is in real estate, or is that just because the asset class is so big that you can't possibly own it all?

It's just so sexy, you know, everyone like parking.

Yes, yes.

You know, it's like, no, it's like the last bastion of non institutionalized real estate in the United States.

Yeah, yeah, you're right.

And it's everywhere. It's 15% of the surface area of our cities. And for us it was the first touch point to your comment earlier. We realized that if we can roll out parking and we can deploy a seamless payment product across parking, then we can move into gas stations and car wash and quick serve retail and then we can move past the car and into the store.

Yeah. Can you talk about the kind of longer term vision for city parking? I feel like there's, with autonomous cars, people are excited about maybe reclaiming some green space. Whenever I see, see a big city and there's just like one flat parking lot, I'm like, that would be way more efficient if it was like three stories of underground parking and then a massive building on top with some mixed.

Retail and some on the autonomous cars, people like to say, oh, when we have autonomous cars, we'll need less parking in cities. And I can see that in some way. But at the same time, I'm not excited to spend every day in Waymos, effectively taxis. I am excited to have my own autonomous car. And the idea that I would drive into the city and then my car would just be driving around all day, like it's got to go somewhere while I'm, while I'm podcasting. Right. Like it can't, you know, so, so it's going to be parked somewhere. And so I think this sense that autonomous driving is going to eliminate the need for, for parking is, is. Or certain, you know, some, some of the parking in cities is maybe off base.

Look, I think it's an interesting question. I mean, when we founded the company, we spent a lot of time thinking about the future of autonomous vehicles. And to your point, cars are not going to circle the block endlessly looking for their next job. They're going to get off the road and back onto the road as quickly as possible. And that's the conversion of parking from parking lot to mobility hub, you know, where you can facilitate the cleaning, servicing, charging, and deploying of vehicles. I mean, at this point, our underlying technology enables kind of seamless, I would say, integration or connectivity between an autonomous vehicle and old world infrastructure. I mean, at this point, we're interacting with millions of Americans every single day that are on our platform or interacting with our locations. And I think at this point we're onboarding 50,000Americans every single day that are signing up for the first time with Metropolis.

Yeah, yeah. The idea, you know, you're not going to. A Waymo is not going to have like a humanoid robot in it that presses the parking ticket button when it's coming in and then is like feeding the ticket back in. It obviously would just happen, you know, if Metropolis can be that hub for autonomous vehicles, that. It makes a ton of sense.

Yeah. Talk to me about the, like the evolution of financing. You're obviously using a lot of venture capital, but I imagine that it's a certain point you need to interface with private equity firms or banks or debt providers. How can you get creative with the, with the different structuring financially on some of these larger deals? And can you give me kind of like a 101 on what it takes to get a billion dollar deal done these days?

Yeah, I mean, we did, you know, the last deal was a $1.8 billion Series C. Amazing.

Huge. Huge.

Yeah.

Listen, we took an old world, you know, 100 year old company that was publicly traded private. Yeah, and yeah, look, it's, if you look at our cap table, it's exactly that. It's venture and private equity in the same rounds. So last round was Eldridge Vista 3L, Tomasek Vista BDT, MSD. And you normally don't find all of those players in one party round, but yeah, look, I think that you guys talk about this more than I do, but the capital markets operate very clearly in these very set boxes. It's credit, it's private equity, it's growth. And we kind of broke through that with an entirely different strategy which is how do you take kind of next generation technology and artificial intelligence and start acquiring old world businesses to scale into the market even faster? But in the context to your question, in the context of our business, it's structured to a great extent like any series C venture company would be structured.

Talk to me about what the communications challenge of taking over a 100-year-old company. It feels like you're fortunate that you don't have the, the stench of maybe like, oh, we're going to buy and lay everyone off. But it is change management is, it is a new structure in the organization. I'm sure there's some communication that you have to do when you, when you buy a company. What has that been like and what are kind of the best practices?

Yeah, it's a great question. I mean, look, I think that we went from a 200 person company organically to a 2000 person company with our first inorganic acquisition and then to a 23,000 person company with employees in 400 cities. So yeah, there's massive change management, there's massive internal communication protocols. But I'd say first and foremost it's like how do we put our employees and our team first? And you're right, we were lucky because this was not a turnaround. This was not a structure where we're looking for cost. Synergy, Metropolis. And this strategy of taking these old world companies private was all about revenue synergy. It was how can we drive more revenue and more value not only to real estate owners but, but to our employees and our team members. So it's been exciting. And what we found is it's really interesting. You see cultural friction on both sides. You see friction in the engineer that's worked at Amazon for 10 years that now works for Metropolis. And then you see friction on the parking attendant. But you find people that are really excited about building a hyperscaler and really excited about how they can leverage and build on the existing infrastructure of this hundred year old company.

That's great. Well, thank you so much for stopping by. This was great.

Come back on when you're. When you have a good reason for us to hit the size gong.

Yeah, we'd love to.

I'm sure you'll have many more in the near future. This is great.

Looking forward to it.

Thanks for taking the time, guys.

Talk to you soon. Cheers. Bye. Next up, we have Andrew Huberman joining the stream. We'll bring him in in just a second. We have some breaking news too. Scott Wu has posted about the new model from OpenAI03. Price drop makes it 15 times cheaper than GPT432K, the state of the art model from two years ago. Meanwhile, the number of use cases is probably up 1 millionx. Kudos to the OpenAI team for dropping the price. So congratulations to everyone over there and welcome to the stream. Andrew, how are you doing? Great.

Great to see you. Thanks so much for joining.

Yeah. What's new in your world?

What's new? Goodness. This week we have a big episode of the Huberman Lab podcast out with the current NIH director, Dr. Jay Bhattacharya, who's a MD and a PhD. He's a unique background because he has a background in medicine, obviously the md, but he also has a background in economics, incidentally, did his undergraduate master's, PhD and medical school training and then was a professor of medicine at Stanford.

Stanford.

He's Stanford the whole way. And as NIH director, he holds a ton of power over what happens for the future of basic and clinical research. And he looks at all of that through the lens of an economist, but mostly through the lens of a public health official. So it's a very unique perspective and folks on X will recognize Jay Vocal. I don't want to speak for him. And label, we have to be careful with labels. But anti lockdown, I don't know. Not a huge fan of the lockdowns for most people. He does say that there are certain populations that he felt should have been kept indoors, but other folks probably in his view should have had more actions.

Well, it's a long episode. I wanted to basically get a preview of it. You guys focus on a couple things. One, how to fix the issues with science and two, the importance of funding research. Both, sorry, basic research and then applied research. So why don't we kind of COVID.

A couple of those different. My big question was like the narrative right now is that research funding is being cut immensely and I was wondering if you got a feel for how severe the cuts are. Is the narrative overblown? What is the balance of cuts between applied and basic research? And then we can kind of go into some of the implications of, of that.

Sure. Okay, so there is an, an upcoming vote in Congress in September, I believe, and it's on the table to cut the overall budget for research for NIH by 40%. 4, 0.

That's huge.

Which is huge.

That's huge.

The budget for research at nih. There are multiple dimensions to NIH and I don't want to get lost in the weeds of it. It's a, you know, NIH is something I'm very familiar with. I was on grants panels for over a decade. My lab was funded by nih. We could go really deep into the weeds. But let's just keep it pretty simple. NIH funds basic research, which is research that it's not specifically geared toward understanding or trying to solve a particular cure or treatment for a disease. So think. Everything we understand about cell biology, not everything, but much of what we understand about cell biology was because of NIH funded research into the functioning of the cell over the course of, you know, 50 or more years. And that led to important implications for treatments and cures for cancer. We don't have a quote unquote cure for all cancers, but many cancers now can be cured.

To be clear, these are non commercial. This is like non commercial research. So things that we cover a lot of.

Yeah, we cover a lot of, cannot be patented. So basic research is, you know, a laboratory wants to understand how cells work, how neurons work, how the immune system functions, gut brain axis studies on how stress impacts health, sleep, et cetera, then the applied work is also funded by nih. So clinical trials are funded by nih. This is an enormous portion of the overall budget. The exact division between basic and clinical trial funding is hard to demarcate. But let's just for sake of this conversation, just agree. Because it's true that most of the basic research and clinical trials that are run in the United States are funded by the nih. The NSF is a separate entity, right? But nih, its specific goal is to improve the health and longevity of American citizens and by extension the rest of the world. Right? Because it is fair to say that while there are excellent research funding bodies in the uk, in Germany and Switzerland and all over Asia and around the world, that the nih, when people say it's the crown jewel, it's the one that's devoted the most billions of dollars to basic and applied research. And it's no coincidence that the majority of Nobel Prizes in physiology and medicine and chemistry and related fields have come from work that was either initially ceded by or certainly supported by the nih. So you can't overstate the importance of NIH funding basic and applied research. That overall budget is facing a 40% cut in September. Now, you know I'm hearing two things out there, right? I wear many hats. One is my lab ran on NIH money. It no longer depends on NIH money. I'm a podcaster, so I have the ear of folks who are like, this is really scary, right? It gets very political because the current administration seems to be taking a more like they're going to revise the way that NIH is structured, potentially 20 plus institutes down to 8. So all of it looks like downsizing. There are a lot of people who are terrified about this. Okay, there's another camp that I hear from a lot, and I can't even say particularly on X. And I want to be very clear, this camp doesn't always lean right. It's pretty even across the board that are saying, wait a second, why are we giving so much money to these universities to do research from our tax dollars? And I know this will upset people who are very science minded, as I am, who care about science, but they're saying, why are we doing this? Right? Some of these universities, not all, but some of them, Stanford, Harvard, Yale, Princeton, ut, Austin, et cetera, the private universities often, although some public ones too, have very large endowments. And they're thinking, why are we funding so much of this work? Maybe these tax dollars should go elsewhere. There's another key issue, and this came up during the discussion with Dr. Bhattacharya, which is there are many, many people, okay, I'm just who voiced to me that they don't want to give their tax dollars to basic or applied research at all. Okay, this is a large and growing crowd because they feel that there needs to be, for lack of a better way to put it, some truth and reconciliation, right? They want two things acknowledged. And I've actually heard this from many of the. Let's just say that the highly recognizable names in the world of Silicon Valley super tech or founders and funders and investors, et cetera, and those two things are the following. One, they want the NIH to acknowledge, or CDC and or others in government to acknowledge that there were in their mind failures during the pandemic, in particular lockdowns that impacted the non laptop class. Okay? We're talking about janitors, staff, teachers, kids, et cetera, that basically had to halt their work and their income. They're pissed off about this, right. That no one kind of acknowledged this. The other thing that they're very angry about is the lack of acknowledgment from the scientific community that the science community makes errors, sometimes errors that for instance, in the field of Alzheimer's research, I want to point out not all the work in the field of Alzheimer's is bad. Much of it is very solid or excellent. But they are frustrated by some recent kind of unveiling of the fact that there were findings that later were found to be fraudulent, basically, and that it was never checked up on. And then that opens up a whole discussion, which I also discussed with Dr. Bhattachary about what's being done to solve the so called replication crisis. So they're upset about the public messaging around health and science. Right. They would have preferred, it sounds like that people in government say, hey, listen, we have a virus that we don't understand and we have ideas about what might control it, but we don't fully understand this. So it was very iterative in a lot of, of people are pissed off. Kind of like when you're, you know, 15 or 16 and you're being told you can't stay out late and then your parents are staying up late, or you know, you've. Or like we had governors who were saying you had to wear a mask. But then we're taped in fancy restaurants on the northern coast and, you know, restaurants and like much in the way that a teenager goes, wait a second, like you smoke pot in college and you're telling me not to smoke pot like this. There's a logical flaw here and we at least need to talk about it. So there's this kind of notion that this isn't my stance necessarily, I'd be happy to share my stance, but that the scientific community has kind of cloaked its errors. And so there are a lot of people in the general public, they're like, don't give these universities a dime, let them dip into their endowments. And the last thing I want to say about that is not every university has large endowments. Most public universities do not have large endowments. It's also true that universities don't like to spend their endowments. And the typical way that they hide behind their endowments or endowment spending is to say that money is earmarked for other things. Right? So which is not to say that it isn't. And Then the last thing, and I know I'm going kind of fire hose here, but I want to make sure this comes out. Is that the big issue that is really on the table as well is this notion of indirect costs. You guys are finance guys. So basically a laboratory might get a grant of a million dollars across four years. So 250 a year for four years and the university then gets what are called indirect costs. They get, let's say 500k. Typically it's anywhere from 300k on the million to up to 750k on the million. There are a few cases of even more than that. The so called indirect funds that take care of administrative costs and the basic costs of doing research. And earlier this year the Trump administration said we're cutting that to 15% across the board for all universities. And that was very prominent on X because Elon retweeted it and a number of other people retweeted it. And I would say while the indirects have been very controversial, I just personally, it's my view, this is my personal view, is that a severe cut to the indirects, while on the face of it might sound good, that's going to disproportionately hurt the public schools without large endowments because they don't have money to dip into. And you know, something more in the range of 30% seemed reasonable to me based on my understanding of what those funds are used for. But historically there have been some challenges with indirects. You know, there were universities I won't mention which caught for spending some indirects on things that were unrelated to the science. You know, perhaps keeping the lights on in the English department, that's typically not the case now. It's for disposal of radioactive waste, it's for, yes, janitors, it's for painting the walls of the building, but it's not for making, you know, a lavish lifestyle for the administrators. Right. If administrators have a lavish lifestyle, it's presumably through some other mechanism. Not idc.

Yeah, there was some example, I don't know if it was UCSF or some school where they had like a multi hundred million dollar administrative budget. And I think people really wanted to push on that and, but I think your concern is like, obviously there's like this massive trust issue between the public and new parts of the administration and science. And how do we rebuild that trust without destroying the entire system? Right. Is that, is that kind of.

You nailed it. And I would like to just highlight in the backdrop of all of this there is one very major emotional Issue that I'm going to catch a lot of heat for this, but I don't really care anymore. And we're on X. So, yeah, a lot of. I put out a post recently that said, hey, listen, I'm very concerned about this 40% overall cut as the most severe damage to science. We're not talking about allocation, we're not talking about indirects. We're not talking about what's going to happen with that body of money except this 40% cut. And I'm very concerned about this. But I don't want to talk about, for instance, the fact that federal funding to Harvard or to Columbia has been frozen. And a lot of people said, well, why not? I have a lot of friends at Harvard Med and at Columbia and elsewhere. And they're like, wait, this is why. Why won't you talk about that? Well, that's actually a different issue. Right. The reason that money has been frozen is that in the eyes of the administration, they are. Harvard and Columbia are violating civil rights laws. Until that's resolved, there's no discussion about NIH to be had. That's not an NIH issue. It funnels through NIH money.

People are tying the issue, even though just because it has some of the same buzzwords. That's interesting. I have a question about the 40%. I mean, there is a world where what becomes important is the ranking of what gets cut. And so my question is, within the basic research that you're seeing, what should we be fighting for, hardest to keep? What basic science or applied science research are you most optimistic about going forward? Because we saw this with GLP1s. I mean, I'm sure CRISPR these stuff. This stuff. Was NIH funded at one point. There was a ton of research. And then we get this big boom, like. But it's 10 years later. What's on the frontier? What are you most excited about and what should we be fighting for? Let's keep the funding there no matter what.

Yeah, great, great question. In fact, the most important question. So regardless of whether or not this 40% cut happens, if it's less, if it's kept the same. A couple of things. First of all, despite the fact that NIH funds basic research and that many laboratories, including my own, for many years focused on basic research, like how does the visual system work? We always had an eye. Every laboratory has an eye toward what the translational implications could be. So we were actively involved in trying to solve blindness due to glaucoma, even though we were focusing on some basic questions. So I think it's a statistical issue. If you step back and you say, okay, in the field of, let's just say, vision, my former field, what is the leading cause of blindness? Cataract. But you know what? You can fix cataract. You can slide out the lens and you put in a new lens. That's being done now. Great. What's the second leading cause of blindness? More than 70 million people worldwide? Glaucoma. You can test for eye pressures, but once the cells in the eyes start deteriorating, people go blind. There's no recovering those cells. So you could say, wow. So, okay, so let's take the top. Let's take the statistically the top causes of blindness, and let's fund those, okay? Retinitis pigmentosa, et cetera. You could do this for pretty much any field. So in the field of brain health, you'd say, okay, dementia, you'd say major depression, you'd say a stroke, you'd say, I'm going to piss off some people because I'm not going to mention their particular suffering here. But it's not hard to find Parkinson's, right? Ms, I mean, it's not hard to know what the problems are to tackle. The problem is deciding what the targets are. And as you mentioned, like with crispr, I mean, in many ways, it was a fortuitous thing, right? Doudna lab was working on bacteria. I mean, it's very hard to predict what basic research is going to lead into these different areas, but we know what the critical areas are. So I'm not trying to dodge your question. I do want to say that the emphasis from Maha on chronic disease and chronic health issues, I think has concerned some of the people in the scientific community. Because, listen, I, as a podcaster that covers science and health, will be the first to say that if you're not sleeping well, your mental health and physical health is not going to be good. You can miss a few nights sleep. But if you're chronically sleep deprived, you're sicker than you normally would be. If your gut health isn't proper, if you're not getting exercise, if you're not getting sunlight, if you're not doing these things, of course, right? But I think that people who are in the kind of hardcore mechanistic science community, like, great, we're all going to strive to do those things, and we should, but not at the exclusion of figuring out signaling pathways that are vital for, like, the GLP1s are a beautiful example, right? This peptide acts at the level of the brain and the gut to make people feel more full. It also, by the way, shut down the debate as to why people are fat. It's because they eat too much, right? They eat more than they burn. Remember people used to argue about it?

Yeah, yeah, totally.

No, no, GLP1 silenced that. Yeah, it's really true now. Are they eating too much because of a hormone issue related to having too much fat? Maybe there's some depression, sure. Like I fully acknowledge, like I'm not trying to be completely insensitive, but like, but we, we isolated the problem by understanding that this peptide discovered in heal. The discovery of this is worth spending two sentences on.

Please.

Gila monster, a reptile that doesn't need to eat very often makes a lot of this peptide. Peptide that it turns out is manufactured by humans too. And if you increase in humans, you know, you're not hungry anymore. And so I think that, you know, the guy studying Hela Monster, the Gila Monster, I doubt was thinking about curing obesity, but there you go now. And of course there's a debate to every one of these things. No medication should ever replace lifestyle factors. Right? Yeah, there are many. I think we sometimes fail as healthy, fit people. We fail to understand that lifestyle factors are very hard for people to implement even if they have copious amounts of disposable income and some free time. It's just tough. It's just tough. It's hard to do. And so I think that the, the emphasis on nutrition and exercise is wonderful. I think, I don't think anyone on either side of the political debate would argue that Maha in principle make America healthy again isn't a great thing. I think what they probably would appreciate hearing now, I'm speaking for kind of the more science, mechanistic, biology minded folks, they'd probably like to hear a little bit more about, hey, you know, they're probably signaling pathways in the brain that are relevant to, you know, to depression. That maybe a next class of antidepressant drugs would probably be good. I don't think anyone would say there's no need to develop antidepressant drugs. We probably take too many of the ones that don't work and have too many side effects as a country. But so when you say what are the most important things? What we really need is a very rationally grounded planning committee to sit back and say, you know what? We're going to put X number of dollars towards this, X number of dollars towards that, X number of dollars towards this. And this is where I think Dr. Bhattacharya really shined on the podcast, he said, and I would totally agree. And I got, I have already. My phone's been blowing up with some anger from colleagues about this. Much of the work that's funded by nih, here's the dirty secret is already completed because they tend to fund things that are very certain to be completed. And so people put forward grants based on work and preliminary data showing I can do this and it's already kind of worked out. Then they use the money for the next iteration. This is the dirty secret of every NIH funded lab. And it tends to favor a kind of pedestrian, more pedestrian science. Then they use foundation money and they use other sources to kind of do the high risk stuff. But if you were to look at like one of the more impressive funding bodies in science like the Howard Hughes Medical Institute, right. Gives the equivalent HHMI folks, as they're called, never. They always say it's not that much money. But let me many. We can do the test, we can take away their money and we can see how well they do or don't do. They don't like doing that.

Yeah.

It's like being on academic steroids, right?

Yeah.

Equivalent of three grants per year, basically. NIH grants per year. And more money translates to. When it's excess money. I know that word, excess money scares investors, especially if they're taxpayers. But if that money is being used to fund really like really bold hypotheses, like it's my opinion that every single peptide, not just GLP1, every single peptide, shouldn't be getting tested in the gray market of like gyms and like biohacking. That's frankly, it's bullshit.

Yeah. And isn't the reason that we don't have good studies on a lot of peptides is it's not profitable to study them because they're effectively naturally occurring. And you can't patent like correct there like BPC157.

Right. You hear about that all the time is will it accelerate healing? You know, the animal studies studies are solid. There's no human studies. My good friend Peter Attia is like, I'm not going to touch that stuff. There's no clinical data in humans. And the rest of us are like, well, I take it and it works. And there's, there's no reason for a, for a drug company to take it. I'll take it transiently for like a, like a joint.

Yeah. But the core issue and I, yeah, I've, I've, I've taken it too. I've benefited from It. I recommended it to John a couple weeks ago. He had a. He had an injury. I just said, hey, I'm not going to like. This is not something I'm going to take, you know, weekly for my entire life. But in certain instances. But. But break down that specifically that you're saying because it's. Because it's effectively science. There's not. There's no real ability. Would a drug company have to create some variation of it in order to patent it? Is that. Is that.

Yeah. In fact, and here I'm going to catch heat from everybody. But, you know, the pharma companies that make GLP1s, I mean, most people have realized that GLP1s are very expensive. And the dosages that they're typically prescribed, people get some discomfort. And so the new thing is people get it compounded at a compounding pharmacy and they're microdosing GLP1s. The number of women that I know who are microdosing GLP1 in service to the spring and summer. Outrageous, right?

Yeah, outrageous.

Yeah, it's outrageous. And the idea is that it's. GLP1 is a good example. I mean, again, insulin's a peptide, you know, but there are many peptides. Like, for instance, I'm very excited about this peptide, Pinealin for sleep. Right. And it's for some other properties, too. But there's no reason why a drug company would go patent Pinealin. They're busy patenting drugs like the recent class of drugs. The. This is an interesting story. Work on a peptide called hypocretin. This peptide is involved in generating wakefulness. It's also in the feeding pathway. And there's a new class of sleep drugs that suppresses the wakefulness pathway as opposed to making more sleepy. So lower abuse potential. The Doras have been released. Right. They maintain the architecture of sleep pretty well compared to other sleep drugs like Ambien, et cetera. Here's the issue. They're about $325 a month. So for a lot of people, that's prohibitively expensive. And they're patented.

Yeah, they put it in context. That's. That's a luxury gym membership.

You know, Pinealin probably cost you. And I'm not suggesting people take Pinealin. I'm not a physician, so talk to your physician and then ignore him or her if you choose. Don't. Don't do what? Don't do that on my. On my suggestion. But, you know, it cost you about 15amonth. Right. And you know, it's. It's for, for me personally has allowed me to. I get two and a half hours of REM sleep a night in a six hour sleep bout. So I've shortened my sleep bout to about six hours a night with a lot of deep sleep. And the Pinealin, is it completely safe? Reasonably, but we don't know. We don't have clinical trials Now. I'm willing to run that experiment on me, but are you willing to run that experiment on you is the question. And that's where it gets down to personal freedoms. And that's a, you know, I don't want to cover too many things here, but the mop. I will say that Robert Kennedy has been pretty vocal about at least before the election about wanting things like peptides and stem cells and supplements to be more widely available. The truth is they're pretty widely available now. And I will put an asterisk on stem cells. I know a very prominent public facing physician who is almost paralyzed permanently from a stem cell injection into the disc of his back. I talked to a neurosurgeon friend who ultimately saved his life by the way. This is a well known story within the wellness community and he said, yeah, you know, the discs can't accept stem cell injections. But if you go down to Mexico or you go out of country and you say, you know my back is hurting, they'll inject stem cells into the disc of your back. So is this to say that stem cells are bad? No, they're not ready for clinical use yet. And the United States has been very careful about things like this because before they were careful, there was a clinic in down in Florida that injected some stem cells into the eyes of people with macular degeneration who were worried about going blind. And guess what, they all went blind right away permanently. So I do want any discussion about peptides and these kind of things. When you get into the realm of stem cells, it starts getting to be a serious matter. So I don't want to give the impression, I'm like, oh yeah, try this, do that. You know, shoulder hurt, put some stem cells in. When you're talking about stem cells, you're talking about cells that can become essentially anything, including tumors.

Yeah, talk about, I'm curious, I mean you and Rob, I feel like, do such a good job of threading the sort of needle around, having conversations without giving endorsements for specific things that are untested and really focused on understanding, helping people understand the world, the science themselves, and then take the actions that they themselves as well as you know, their doctors believe they should take. Do you think that podcasters generally in health don't really understand broadly the responsibility they have? I remember when I was in college, there was a popular tech podcaster at the time that would pretty much widely endorse microdosing, right? And I looked up to this guy, right?

Psychedelics, right?

Yeah, psychedelics, right. And in hindsight, I look back on that and I'm like, that is so wildly irresponsible to basically recommend, you know, that there's microdosing, some people report tremendous benefits, but to sort of broadly endorse something in hindsight feels insane. And I don't think people fully grasp the responsibility that they have and just how much impact they can have on an individual or how a community thinks about something.

Yeah, I mean, Rob and I, we take it super seriously. I mean, I've probably taken, I've taken heat for a bunch of things in the larger science community, but probably one of the things I've taken the most heat for is my belief that certain supplements can be helpful, not as a replacement for, you know, good behaviors and for prescription drugs. I'm a fan of certain prescription drugs. Listen, I'd read an episode on ADHD and talked about the prescription drugs and took a ton of heat from the kind of natural folks. And then I did one on behavioral tools for ADHD and took a lot of heat from the folks who said, hey, listen, my kid got a lot of benefit from taking stimulants for adhd. So we, we do weave back and forth and on that kind of knife edge. And it can be tricky in terms of microdosing because I saw it come up in the comments before today's discussion, and so it's just kind of flagged in my mind. The data on microdosing psilocybin are basically clear that it's not known to have any major effect on major depression or some of the other things that psilocybin is being tested for clinically, including major depression, ptsd, et cetera. The data from clinical trials out of Robin Cardart Harris's lab at UCSF and elsewhere on high dose psilocybin. So macro dose, right? Two and a half grams, four grams is the, quote, unquote, heroic dose. If you're speaking Northern California language. Those dosages done in pretty maybe two weeks apart or so, with the support of a clinical staff going into it, through it, out of it, et cetera, have been very successful. In those trials for the treatment of depression, microdosing has shown very little effect. MDMA was, was put up to the FDA last year as a potential treatment for ptsd. It has not been approved. It did not pass approval. Unfortunately, as a component of those trials, there were some sexual improprieties in one of the clinical conditions that, you know, all it takes is, you know, one bad incident, right? And then it was very. But it cued people to this larger theme. How are you going to protect patients who are really, they're not incapacitated, but they're not in a position to really advocate for themselves that you need probably multiple clinicians in the room and checks and balances. So I think ultimately it will be approved. The remission rates on ptsd, by the way, from properly dosed and spaced MDMA is remarkable. Up to somewhere between 60 and 70% remission rates on PTSD. And the new thing that people are very excited about is ibogaine or iboga 22 hour psychedelic journey. This has been tested largely from Nolan Williams lab at Stanford. Seems to in one or two sessions. It has led to a significant number of veterans essentially ceasing their alcohol use disorder, what is typically called alcoholism and opioid use disorder. But you need to be heart rate monitored while you do this. The point here is that any discussion that we have on the podcast is, as you can tell, like, I'm not known for being very succinct. It's. And it's for a reason, right? You have to flesh out the conversation around these things. You can't just say, psilocybin is good. Mdma, yeah, like, works great. Remission, you know, people can really get hurt. And I think discussions around whether or not you take magnesium threonate or bisglycinate before sleep, like, okay, you can probably speed through those a little more quickly. But when you get down to things about psychoactive drugs, in particular Schedule 1 illegal psychoactive drugs, when you get down to things about hormone therapies, you know, you're starting to get into the, into the realm of where people can really screw themselves up. That said, there's, you know, there are a number of things that I hope that the NIH and the, and the public health messaging going forward will be more expansive about, which are drugs that are currently prescribed at enormous rates in the United States and elsewhere for which there are very severe side effects. Like there's this whole thing about people taking finasteride and dutasteride men to keep their hair and getting permanent sexual dysfunction, right? Permanent sexual dysfunction. Some of them killing themselves as a consequence of this. These are young guys. It's clear those drugs have very different effects in young people versus old. Some people can use them safely, some people can't. These are FDA approved drugs as well as I think we're finally coming around to the idea that the SSRIs can be very useful. Useful for things like OCD in certain cases for depression, but it's a double edged blade. A lot of people suffered it as a consequence. So I hope very strongly for three things. One, as a scientist who essentially has my podcasting job because of the way the NIH supported me coming up through graduate school, postdoc, my lab, I think a 40% cut would be too severe. Regardless of the IDC issue. I think it's just too severe. It's going to kneecap science throughout the world and health the growth of new discoveries throughout the world. I'd like to keep that funding level, I really would. What happens to idc? What happens to Harvard and Columbia? Separate matter within there. But then I would very much like a very thoughtful committee to go in and look at every single grant. Every single grant by area.

You have to have a scalpel, not a couple.

And you could say, I mean listen, in the field of neuroscience, am I the most qualified to decide what grant should be funded or not funded? No. But I can tell you that I sat on study section for a long time and I can tell you there were clusters of grants that didn't solve what we call what good labs call the deletion test. If this laboratory didn't exist, would it change the direction of science? You have to pass the deletion test. Solve. You have to pass the deletion test. If you didn't exist, would it matter? That's a harsh test. Nobody wants to be subject to that test. But anyone that's getting millions and millions of taxpayer dollars, and by the way, millions of taxpayer dollars often to work on and kill. And we could argue about that. I think most people are species. They'd rather see a mouse or a rat life or a non human primate life. But you know, we're killing animals. We're occupying the lives of graduate students and postdocs with taxpayer dollars. The work needs to be justified by passing the deletion test. And the idea that AI could do that, maybe. But the idea that a jury of close peers are going to decide the deletion test and if you pass, that's flawed in my opinion because the in culture, we don't have time for this. But the in culture of science is, you know, people aren't like helping each other out, just to help each other out. But when you're very close to something, it's very hard to make a really clear decision about it. And when you're too far from it, you're not qualified to. So we need people that are in the middle who can really go in and say, okay, in the field of, let's say, visual repair and neuroscience, give me a scalpel. I hate to do it, but also some are going to pass. And then I do think that instead of cutting people's funding entirely, there should be initiatives saying, hey, listen, if you want to run a laboratory, would you be interested in working on these important projects that maybe the Americans paying for this research ought to be able to vote upvote. Why not? I mean, it's 2025 and the NIH has done beautiful work over the last. Funded beautiful work over the last hundred years. But I think an update in the way these things are handled is great. As I say that every single one of my colleagues is probably quaking that they're not going to pass the deletion test, except the ones that know they would absolutely pass the deletion test. And then the question is, who's making the decision? And I'm not saying I should be making that decision, although I have some strong opinions about what's great, what's meh and, and what's like, lousy. But I do think we need, we need to be more discerning. The same way that a bunch of VCs would sit around and say, like, hey, what are we getting for this investment?

Well, yeah, I think the question they ask is, if we don't make this investment, is the, is the world going to be, you know, worse off? Like, does this company, you know, I think it's not quite the same test, but I think it's important for investors to run that, that process and understand, are we really funding the future that we want or is this kind of a rounding error and the dollars be better elsewhere?

Can you give us a white pill like, are there any other organizations that can come in and help? I always think of what happened in computer science and artificial intelligence. Google created the transformer architecture. Not patentable. It created large language models and we got ChatGPT and we got all this great stuff. It would be amazing if we could find a way that the big pharma companies that are big and profitable wind up stepping in and funding some of the gap. Maybe there's nonprofits, maybe we just, you know, the sheer will of the American people, we wind up voting for more funding. But how do we really make science amazing and ensure really strong progression over the next few decades?

Yeah, I mean it's, that's the, the key question, right? So I mean, listen, I grew. I was born at Stanford Hospital, you know, did my training, much of it at Stanford. I'd say I'll probably die at Stanford, hopefully in the hospital. I'd like to, I don't know, died getting morning sunlight or something. But my dad's, my dad's a physicist turned computer scientist. Right. Worked at Xerox park, which was an incubator for ideas. Right. For many years. So I grew up in that landscape. And you know, he always said, you know, eventually degrees in computer science science will be important for starting companies. Right. But when I was young, that wasn't the case. Right. People dorked around on computers and made video games. And then we saw cell biology migrate into the world of biotech. And for the treatment of certain diseases. It has been. We can't forget that there are certain basic discoveries going back to our earlier discussion that have led to tremendous treatments like the drops to lower eye pressure and glaucoma. If you catch your eye pressure elevation early and you take those drops, you will keep your vision.

Wow.

Right?

You'll keep your vision. So it's not like we can't cure glaucoma, but you have to. That's based on basic work. So the real key here is to ask, you know, so where has it been done successfully before? It's been done successfully in, in the world of computer science and AI. Right. Growing up I heard about AI. It was kind of like, like no one, no one took AI seriously growing up or brain machine interface was like sci fi.

That was so sci fi seriously.

Now, my good friend from, from childhood, Eddie Chang is the chair of neurosurgery at ucsf. I mean, he's using bmi, brain machine interface and AI to get people with locked in syndrome to, to speak and to. And Elon's going to get them to move. I fully believe that neuralink is going to solve paralysis. Really do. I'm not saying that because we're on X. I know the head neurosurgeon there. They're well on their way. Right?

Yeah.

So the question is when should it move into biotech? And you know, Peter Thiel had the brake breakthrough labs idea. He was paying people to not go pursue graduate degrees in labs to get the PI like me papers and advance their careers and then maybe go get their own lab. But to take Great ideas and go through breakthrough labs. I don't know what happened to breakthrough labs, but I do think that incubators like that, where you have enough money to test, you know, to move fast, right? You know, move fast, break things, model, figure out whether or not something is possible, promising, and then advance that. And here's the key. As soon as something looks promising, you want to be able to pour human power onto that. And the problem in academia is every graduate student and postdoc needs a first author paper in order to have the potential to get a job. And so one thing that I'm, I've, you know, been in Jay Bhattacharya's ear about is, you know, I think, I think the independent investigator model of science in this country where it's the Huberman Lab or it's the whatever lab named after the PI, I think that's a flawed model. This is going to really upset some people. There should be labs named after a particular mission, like Curing Blindness laboratory or Solving Alzheimer's laboratory. And then people can collaborate within that lab without the idea that they're necessarily going to go start their own lab. In Europe, typically there's a lab head and it's very hierarchical. But people who get PhDs off stay in those labs as permanent careers. They get paid better and better. The NIH has not had a mode to keep paying people as staff scientists. And most people don't want to go off and run their own lab. So I think the ability to iterate more quickly, much like a tech company funded by nih, would be enormously beneficial. And I could be wrong, but I think it's going to be music to most people who would like to go to graduate school. Because what it means is that you can be on the bench next to someone and they get the thing that's really interesting. And rather than feel like shit, I got nothing, you can now start working together to accelerate the progress of that work. And so you now who can go off and run a lab becomes a separate issue, but it really becomes discovery based science. And we can't forget that the taxpayers are funding this. Right? So this is like as much as we'd all like the next generation of scientists to all have their own labs. I think what we really want is for the next generation of scientists to have the opportunity, opportunity to make fundamental discoveries that seed health and cures and treatments for disease, right?

Yeah.

I mean, do we really care about the careers of scientists? No, that we care about treatments for cure, treatments and cures for disease. The careerist model.

Well, I do think, I do think it matters that young people aspire to be scientists. Yes. And that is a concern that I have about the destruction of trust between the public and the science community. Via Covid. The concern is that you'll have a 10 year period, a decade, a decade where young people say, well, I'm not going to go into science because they have some. And I think that that's the work that you do with Huberman Lab has potential to heal that divide between the public and the science community. And one of the reasons why I think it's so important.

Yeah, I mean, listen, one of the most important, the three most gratifying things I can hear as feedback for the podcast are, I can't believe this is free. You know, that that feels good that people are sleeping better. Like, oh, goodness, I made a few morning sunlight and dim the lights and a few things in the evening. And like, I'm sleeping better because I know that catalyzes many, many other positive changes. And then the third one and the one that's most heartening to me as a scientist, less as a, you know, a health podcast, is when people say, like, you really turned me on to like, neuroscience or, you know, or I'm going into the field of psychology. I want to be a therapist, but I'm really starting to incorporate some physiological tools with my patients. That to me is, those are the three most gratifying things. You know, I, as you can probably tell, I'm very impassioned by this. Right. Folks close to me know this is how I talk all the time. No, I haven't had any caffeine in the last, you know, four or five hours. Just kind of.

We, we have, we have.

I do, I do.

I'm glad you love this stuff. All right.

We love it. We drink it every day.

We debated. I, we didn't have time, but I debated having you join the call. And it was just stacks of cans so you couldn't even see us.

I do four of these a morning minimum. But I have a very high caffeine tolerance and, you know, fast metabolizer. The fast metabolizer. And, and, you know, there are many issues I care deeply about, but what you just put your finger on is so key. If this 40. I know people are like, don't give them this money. They fucked up during the pandemic. Excuse my language. Or they lied. Listen, these kids that would potentially become scientists, we can harness their energy. They didn't lie. They didn't do anything wrong. And if we train them properly and we put them into an environment where the spirit of that environment is just discovery based, maybe a little less careerist, a little more discovery based, and we really fund the most important work, we can really transform the treatment and cures for all these diseases that everyone is so concerned about. I, I know that because it's proved to be the case every single time. If you look at aids, there are two things that led to the treatments. And you know, AIDS isn't fully cured, but it's a very different landscape now than in the 80s. Two things, right? Money and emotion. And it came emotion first in money and then a bunch of scientists in labs and a bunch of funding. Okay, so where the reason we don't have a cure for schizophrenia, Even though it's 1% of the world's population, is, sadly, it tends to run in families. There's not a strong lobby for research on schizophrenia. The director of NIMH told me that former director Bob desimo, the reason there's so much interest in autism is that any human being with half a heart looks at a kid. And I'm not talking about kids that are on the spectrum or a little neurodivergent who I work with those people, we're not talking about that. I'm talking about kids that have stereotype motion, can't be in an open environment, will forever need support. I'm not trying to split the spectrum, but there are kids on the spectrum that really struggle. And so the reason there's so much interest in finding treatments for things like that or for major depression is because many, many people suffer, right? And what's required is money and scientists working on the right problems and ditching the things that aren't promising. The faster you can iterate, the better. And so I do think the Silicon Valley model has something to offer. And I don't know what the latest update is on Breakthrough Labs and if Peter's moved on from that, I haven't heard much about it. But that doesn't mean that they don't have something interesting. But look, I, yeah, the key is.

Like if, if you have one, you know, one car that, that comes out, it's not so great. You don't destroy the car industry, you know, you don't, you don't like basically dismantle the entire system. This car, you know, hurt people. So we should just eliminate all cars, you know?

Yeah, yeah.

Look, I think in engineering we're on X after all. Like an engineering, more of an engineering stance on, on basic research is Helpful. It's also true that, you know, I can name off countless discoveries in biology where, you know, an accident of leaving something, you know, on the shelf too long or something led to an interesting discovery. I mean, there is some fortune, gratuitous aspect to it. But now with AI, you can also run lots of hypotheses in silico, more in vitro, then move to in vivo models. I mean, there is a way to make the money go further, but ultimately there's no replacement for great ideas and there's no replacement for human energy. And the best source of human energy that I know is youth. And so you need to capture people when they're coming out of college, not send them to work in a lab for three years before they might go to graduate school, then realize that they can't make a good living compared to their friends in finance. Like, you're not going to get rich doing science. Although some scientists do get rich, right? I mean, Genentech was started by scientists. I mean, there are companies, you know, but ultimately you want to harness that, that energy of youth and the spirit of discovery and give people the resources. I had wonderful mentors, and one of the most important things they taught me was you get money, you give it to young people and you get the fuck out of their way and let them try. And then when they don't work, you guide them and you say, listen, you're wasting your time. Move on. Like, no, no, no, no. And they get bound to it. No, move on. It's like a bad relationship. Cut and run and then you just start the next thing and eventually you get hits. And then if you compile human power onto those hits. Right. I'm just describing what everyone knows to be true. But this is not the way that the system has been designed. It is. It is old and clunky and it doesn't need to be scraped or, you know, hacked in half almost at a 40% cut. But it does need to be revised. It really does.

Yeah.

Makes a lot of sense.

Well, this is fantastic.

I want to hit the gong for you. Can you give us a stat? How big is the Huberman, Huberman Lab podcast now? Can you give us an idea of the scale of the impact you've had? We like to hit the gong when we hear.

He's looking over Rob, he's like, no.

No, don't leave anything.

Even some rough.

Just any number. Maybe just one episode. Hit a million downloads. Give us some number, something.

Those are rookie numbers.

Those are rookie numbers.

Yeah.

So.

Well, can I say One thing first, please.

Anything.

You know, rankings reflect acceleration, not absolute reach.

Yes.

Like Rogan. Like, when people go like, Rogan's not in. Rogan's reach is, like, beyond. I know the numbers. I'm not going to talk about his numbers, but it's like. It's an order of magnitude greater than all the other major news hubs combined.

Oh, yeah, yeah.

Unbelievable. But no, we are very blessed to reach, you know, 20 million people.

20 million people. Let's go. Let's hit the gong. Thank you so much for coming. Going on.

This was an important conversation. I'm excited for people to get the. Get into the entire episode.

Yes. Yeah. It's fantastic. It's over four hours long.

No, and thank you.

If nothing else, you know, with the Huberman Lab podcast, I. I always say, if nothing else, I'll cure insomnia. But no, the. You know, it's timestamp, so you can break it down. If you want to get right to the vaccine stuff, skip to that balance.

It's fantastic. I listened to it last night.

Can be digested however you want.

Yeah.

I want to say thanks to you guys. I started watching your show since you had Rob on.

Fantastic.

What you guys do is great. I also think that you're transforming the way that media is. Is, you know, dispersed each week and, you know, and. And it's awesome you guys x doing what you do and elsewhere. So thanks so much.

We appreciate it.

Yeah, you guys are both welcome anytime.

Yeah, anytime.

This is really fun.

Never need more matina. Let us know.

Oh, yeah, we can always come.

We're gonna park a fridge right here.

Here.

We're gonna do it. The guys. The whole. The whole crew back there is cheering.

Where are you guys based?

We're in Los Angeles. In Hollywood.

Yeah.

All right. We come right over.

It's a nice.

We'll see you soon. Thanks so much for hopping on.

Awesome. You're the man. Cheers. All right.

Fantastic episode. Very fun.

Well, folks, we had 30 minutes scheduled with Andrew, and we just.

We went longer, but we have some breaking news. We can go to the breaking news camera.

Hope the breaking news printer.

The breaking news news printer.

Want breaking news.

Watch it live as it comes out. I got a note that it's slightly out.

I see the laser work.

There we go. What is this? What is this?

We got. Whoa.

Shane Co market says x Xai are live on polymarket. When I check polymarket to get a gist of what's going on in the world, I always wish there was more context. Now there is the first of many XAI partners with Polymarket to blend market predictions with X data and grox analysis. Hardcore Truth Engine. See what shapes the world.

The Truth Engine. Let's give it up for Truth Engines.

We love Truth Engines. This has been a fantastic show. We will be at YC Demo day tomorrow. We have a flight to catch. We covered most of the news. You've heard it, we talked about it yesterday. Open. I hit 10 billion in ARR. Almost 2. Yeah, we hardly touched the end of 2024. Yeah, it's just. It's just casual 10 billion ARR. You know, it happened.

Happens, happens all the time.

Yep. And yeah, I think that's pretty much it for the timeline. I don't know if there's anything else you want to cover, but I mean, maybe we can close it out with this fantastic post from Mert, who's been on the show from Helios. He says having $300 million liquid is truly an interesting limbo. Golden handcuffs in a way. You have almost infinite money for daily expenses like groceries, Amazon binges and chill little vacations. But you can't really splurge on anything sick. You have just enough money to think about making a real boy amount, but a razor thin margin for error if you screw it up. You're also at the exact amount where it becomes cozy and hard to motivate yourself to go that much harder. Tough spot. Tbh.

Tough spot.

Oh, he's an all time poster. I love him. We gotta get him back on the show. He's so fun.

Banger.

This is great. Yeah, great place to end with a banger. Anyway, thanks. Looking forward to tomorrow Podcast and Spotify. If you're in SF and you're at demo day, come by and say hi. We'd love to talk to you.

Looking forward to it.

Have a great one, folks. We'll talk to you soon. Bye.