The rise of “workslop,” Alibaba’s insane new deepfake model, Tether’s MASSIVE valuation, and more | E2183

We're on our way to commodification is I think what's happening. This could very much look like storage or compute in a couple of years. Looked, you know, 20 years into Web 2.0, which is people just start stop thinking about it and can't tell the difference. It's completely possible that there'll be 20 different models that all have, you know, 99% similar results. Not for super intelligence and trying to cure cancer or fold proteins or solve mysteries of the universe and, you know, go through every single image on in the world, but be able to tell you how to make sushi rice really well or, you know, where to plan your trip for your family in Europe for two weeks, you know, at this price point. So we're going to see that commodification happen. I think right now. If you were to switch the models from the major companies for 50, 60, 70%, maybe 70% of queries, I don't think consumers could tell the difference.

This week in Startups is brought to you by Northwest Registered Agent. Starting your business should be simple. With Northwest Registered Agent, you can form your entire business Identity in just 10 clicks and 10 minutes. From LLCs to trademarks, domains to custom websites, they've got you covered. Get more privacy, more options and more done. Visit northwestregisteragent.com today. AWS activate. AWS Activate helps startups bring their ideas to life. As you build and scale your business, activate credits grow with you to support your changing needs. Apply to AWS Activate today and receive up to $100,000 in credits. Visit aws.Amazon.com startups credits and Vanta Compliance and security shouldn't be a deal breaker for startups to win new business. Vanta makes it easy for companies to get a SOC 2 report fast. Get $1,000 off for a limited time at vanta.com twist all right, everybody, welcome.

Back to this week in startups. I'm Jason Calacanis with my co hosts. With me today, my co hosts, Alex Wilhelm and Lon Harris.

Hey, everybody.

Hey, hey, hey. How you guys doing?

Let's go to our first story, which in the media space is super interesting. I saw south park had taken a couple of weeks off and they're back.

They took last week off under strange circumstances. The day of the show, there was a tweet from the south park account apparently signed by Matt Stone and Trey Parker saying, apologies. We make the show, you know, right up to the moment that it premieres. That's how they always have made it. So that that's how they get it so timely. And so they said, but this week we just missed our deadline. No south park this week. We'll see you all next week. Now that has happened in the past, but it's rare. And in this charged political climate with the Ellison family having very recently taken over Paramount and now seeking to take over TikTok, Warner Brothers Discovery, they've got all these irons in the fire. There were people suggesting maybe there was some behind the scenes, like did south park get pulled last week? We have no confirmation one way or the other. I want to stress that's pure speculation. But so this week, apparently south park is back. A brand new episode tonight and it's going to deal in some way with prediction markets is what they're saying.

I did notice that online, the description says that there's a prediction markets, that they nail it every single time. And man, they have been absolutely brutal to the 47th administration. It is nuts how brutal they are.

I just want to say that I'm kind of shocked that prediction markets have reached such a level of prominence that they've made it onto South Park. I mean, often we can kind of anticipate what topic they're going to discuss. But I didn't think that polymark and Kalshi were big enough yet to Jason, have this level of cultural importance. Are you surprised? Because I'm kind of blown away.

They have huge cultural cachet. They have been known for taking down any sacred cow. They went after Scientology, they went after Kanye, every president, Biden, Trump, Obama, Clintons. I mean it has been. They are just known for gloves off. And I think it's great that comedians can feel free to go for it. And I saw Jimmy Kimmel's last night. I don't know if you gentlemen saw it as well. I just want to say he did a bang up job and I think he was incredibly sincere. He talked about how he felt terrible about Charlie Kirk being murdered and he talked about being a Christian and being affected by the widow of Charlie Kirk talking about giving forgiveness to Charlie's assassin. So, yeah.

And Jason, that monologue racked up something like 13 million views in 14 hours. That's not the exact latest data, but it had a huge resonant impact on the Internet. So I guess people wanted him to come back. And the question now is just what does the government do in response? And we'll have to see. But it's good to have him back. I'll just say that to keep things neutral. Ish here on the show.

I would note Also, the President went on Truth Social last night and said he's very disappointed that ABC has brought Kimmel back. And he suggested he's going to look for other potential avenues to get him kicked off abc. So for the people who are arguing this had nothing to do with Trump, it was just low ratings. This was not political pressure. The President himself is continuing to apply pressure to Disney.

He.

He's outright stating he wants Jimmy Kimmel gone.

Well, that's not gonna happen. I don't think that's gonna be the way this breaks down. I mean, Brendan Carr, Jason, is someone that you've had on the show.

Yeah. Many times.

You've had good vibes with him. I mean, I've gone through the transcripts. You guys have riffed a lot on Starlink and such. But just because he's been a twist regular, did you foresee his actions that we've now run into because a friend of the show has now become a bit different? I'd say yeah.

I mean, I said I'm disappointed. I think I may have even used my disgratiad for the behavior here. I'm gonna just call balls and strikes. I think when the worst of the 47th administration is when they go vindictive and they go like trying to settle scores or going for the censorship. Both parties have censorship like obsessions. They both have big money spending obsessions and they both have war obsessions. These are the three worst things about these two parties. That's why we need a third party that's not for war, not for censorship and for balancing the budget. But, hey, listen, we don't have to go too far past the mandate here. We've got a big, giant docket to get through.

All right, so I'm going to go ahead and kick us off. Jason, I want to talk about the latest model from Alibaba. It's called Juan. And we've all seen a number of, I'll just say it, very high quality deep fakes spring up around the Internet. People showing off what they've built with it. But I think most interestingly, there's been a explosion of people showing off how they've used this model to not just make a deepfake, but to essentially take their own video and create something very, very, I would say, surprising and compelling. So here is an example that I pulled from Reddit of this in action. If you're on the audio version, it's a man sitting in the car and he's made himself into both Sydney Sweeney and Mark Zuckerberg and And let's roll the clip.

So I'm trying one animated. I'm trying to replace myself with a different character. Take a look at this. So, like, I can create fake ads.

With this if I wanted to.

What are you talking about? What? That's crazy.

It goes on from there. Now, the background here, Jason, is that Alibaba, which makes both the Quinn models and the one family of models, put out something called Juan 2.2 animate 14B. 14 billion parameters is what that means at the end. It came out on September 19th and they said it's, quote, a unified mod for character animation and replacement with holistic movement and expression replication. It's an open source model. And I believe the cow has left the barn when it comes on copyright infringement and deepfakes.

You know, in certain jurisdictions that's gonna be true. You've never been able to have copyright protection, intellectual property protection in places like China, or if you did, it was with a lot of work and a lot of hand wringing. I remember going there and people would chase you down the street as an American with just buckets of DVDs 20 years ago and try to sell you DVDs that were, I think, two renminbi each, which at the time I think was eight renminbi to the dollar. So it was four DVDs for a dollar. It's basically 25 cents, 50 cents depending on the Blu Ray or whatever. And they would just hand you 10 of them at a time. And I bought some of them just as a lark to see what they were like. And it was software as well, like every software program. And, yeah, like bad copies. And this is what you see sold on Canal street in New York back in the day. But we're going to have our work cut out for us with this because I'd say this is 80%, 85% of the way there. And so it's obviously going to get to 100% in the next year or two. And it would be indistinguishable. So you will have to take the approach that if you see somebody saying something crazy and it's not on their official account, it is not them.

I believe we've already sort of hit this point that was purely theoretical up to even a few months ago. There was that when everybody was saying, oh, Trump has been gone for a week, he's disappeared, he's sick, he's dying. Then Trump came out and made that statement and everybody was looking at that statement. There was like he did some weird thing with his thumb or There was like an edit in the video and so everybody was like, oh, there's a glitch. It's AI. That's not really Donald Trump. And it was. I mean, as far as we know, it does seem like that actually was. So I think we're already at the.

Point now it was a splice cut. I mean, if you done video editing, from what I can tell, what they did was. Yes, exactly.

But it would not have been even possible with the AI technology we have today to create a 20 minute believable President Trump speech that only had one little glitch. That was the President. But I just think we've already crossed this Rubicon that if you see a famous person on video saying something, you should automatically now be like, is that really them? Or is this some elaborate deepfake text to video prompt?

All right. It's a classic dilemma. Startup founders and their engineers have tons to do, but without taking time for painstaking compliance processes, they can't start actually winning deals. You don't want your Devs worrying about SOC2, you want them building your product. Of course, it's common sense. That's why it's time to be Vanta, your first security hire. They're the trust management platform that helps you automate your security and compliance, letting you get out there and start winning deals without spending all your time and money on preparing for audits. AI is making it much faster for teams to build products and start shipping. That's awesome. But it also has everyone hyper focused on security. You don't want your team to be left behind. As an investor, this is exactly the kind of product I'm always recommending to founders. Move stuff like security and compliance off your plate with confidence so you can focus on your next customer or hey, your next fundraise. That's why over 10,000 companies, including some of our favorite startups like Rytr and Cursor, are already scaling with Vanta. We're going to help you save $1,000 today through the Vanta for Startups program. That's right, go to vanta.comtwist and save $1,000 for a limited time. That's V A N T A dot comtwist.

Okay, two things about this that really matter. One, the videos that I showed were someone basically transposing someone else over them. A compelling video that's not real time yet. So Jason, I couldn't put myself up as Scarlett Johansson on Zoom. Those videos were then were processed and then synced up after the fact. So it's not real time yet. That provides, I think, some.

That's very important. I thought this was real time and I was absolutely flabbergasted when I saw it. What happened here is he spliced it together after and there was some rendering time. Yes. So these things also are going to cost a pretty penny to do. You know, when you do Grok images, which I think is Nano Banana and Grok are the two best right now. They take time. They take a little bit of time to make these videos and they typically make five second videos. So.

Here'S the thing though. Because it's open source though, Jason, you know, anyone with a couple of GPUs in Iraq can probably run this at a reasonable clip in their own house, is my understanding of this model. 14 billion parameters is not a trillion. It's not the biggest thing out there. So as it gets faster, I mean, that gap's gonna get narrowed down six months, eight months, not far until it's real time.

One more observation here. The guy who made this video is very smartly not turning his head like this. He's looking straight on. That's when this deep fake stuff messes up. If I go like this, it has trouble matching Sydney Sweeney straight onto her profile and you'd get some glitchiness there. But if I look straight into the camera the whole time, even if I move my arms around or I do this, it gets to keep consistent and it looks better. So this guy does understand the limits of the technology he's using and is putting its best foot forward, I think.

Okay, Jason, though, I want to know, does this create the need for a new and more robust online verification system? So that way people who are in fact famous or celebrities or just, you know, well known can have a little modicum of protection around their Persona.

Yeah, my friend Vinnie Lingham had a, a startup, I forgot the name of it right now, where he was using the blockchain to certify this is the person that you're saying it is. And yeah, this is a perfect thing for crypto and distributed computing. If it's not from my Twitter handle, my LinkedIn or a URL I own and control, then it's not me. So there's a lot of different ways to do verification and I think this is going to be a huge opportunity for the platforms. So if, you know, I've been tweeting and Instagramming from the Jason handles on each platform for decades, man, you're going to have a. Those companies are going to have a leg up in becoming a place where you know that if I posted it to my official handle that you know you can trust it. So it's going to be actually big for those companies.

It's called Civic, by the way, Vinnie Lingam's company, he's co founder of sync, a blockchain based identity verification platform.

Yes. I want to make one more point before we move on, which is that there's a pretty big startup angle to this because this is a Alibaba model, it's open source. Great. Love that the technology's out there. But there are startups doing kind of synthetic video generation that are doing quite well. The one that comes to mind is synthesia. They reached 100 ARR this April. And so what I'm going to be curious to see, and I don't actually know where this is going yet, is do these strong open source models from China begin to undercut startups out here in the west that are monetizing technology pretty effectively? Do companies for example, Jason, decide to roll their own video creation products using an open source model and their own hardware? Because Synthesia has to now compete against a potentially free offering. Though of course it does offer enterprise support and so forth. But it's compelling. It's going to be harder, I think, to not say yes to free, even for the, the more conservative companies.

Yeah, so what happens is with free versus paid in, you know, these kind of major markets, we saw it in Gmail, you see it in web hosting. There are ways for you to put up your own WordPress server or you could, there's an open source Slack competitor that you can use and you can pay $0. So right now I spend probably low tens of thousands of dollars using Slack between, you know, two or three different projects maybe. Okay, yeah, if you have 20 people, and I think it's 20 bucks a month. So you know, 400amonth. Yeah, $5,000, let's say $6,000 a year, $7,000 a year. So I could save 7,000 by using one of these other services, popping up my own server. And then you start thinking about that and maintaining it. As long as the price isn't so dramatic, people are going to go with a stable, steady, performant product. And the people who are really cheap probably would never pay for the enterprise versions of Slack. They would look for a free option like Discord. So for a lot of communities they just use Discord for their community like we're using now. And I think you can. Yeah, we'll promote the Discord in a moment. I don't know if I have the URL set up yet, but we've been experimenting with Discord. It's free. So I'm thinking about moving all my founder activity off of Slack because we don't get the whole history and all those features and people seem to like Discord. So for large communities, Slack gets too expensive. You'll move to Discord, but I don't think this will crush Synesthesia or Synthesia.

Synthesia.

Synthesia.

Synthesia, I think will do just fine. The people paying for that want a faster, better product. But it could create downward pricing pressure.

Yes, but it also could create a better product for them. So I was trying to figure out does Synthesia use its own models or does it use third party models, you know, stuff from OpenAI or anthropic, what have you. It turns out it does both. So I'm hoping. My optimistic take here, Jason, is that companies like Synthesia can take what Alibaba has released for free, bake it into what they currently do, and make something better even faster than they could have on their own, and then keep charging for and building great companies because that would be better startups, better outcomes and happier VCs, which would be the best possible outcome, I think, for everyone in the industry.

Yeah, I mean, I think, I think that's what we're seeing. That's what we're seeing with so many of these tools where, you know, they give you the option of, well, which model would you want to use to do this? And I think that's. It kind of goes back to the future that we were talking about so often on the show of like the LLMs. For now, they're, they're the big product that everybody knows, do you use Claude? Do you use GPT5? But eventually that will become the background layer. They're powering the apps that you use that sit on top of those that get to pick and choose whatever LLM they prefer to do the function they're trying to complete.

We're on our way to commodification is I think what's happening. This could very much look like storage or compute in a couple of years. Looked 20 years into Web 2.0, which is people just start stop thinking about it and can't tell the difference. It's completely possible that there'll be 20 different models that all have, you know, 99% similar results, not for super intelligence and trying to cure cancer or fold proteins or solve mysteries of the universe and, you know, go through every single image in the world, but be able to tell you how to make sushi rice really well or you know, where to plan your trip for your family in Europe for two weeks, you know, at this price point. So we're going to see that commodification happen. I think right now, if you were to switch the models from the major companies for 50, 60, 70%, maybe 70% of queries, I don't think consumers could tell the difference.

I was actually reading that, you know, there's all those like character AI, like conversation chatbot systems. I was reading something about those where even they're discovering that different models are good for different kinds of conversations. Like this model is better if you want therapy. This model is better if you wanna be a little sexy flirty. This model is better if you're looking to go real deep on STEM topics or literature. And that's really fascinating. Those models aren't being built with that in mind. It's just some of them work out to be better for different kinds of chats.

So that would be emergent properties. I mean we're literally getting to a science fiction moment where they're starting to have personalities in them. We're all familiar with aws. Amazon Web Services, that's the cloud platform that powers so many of your favorite brands. But do you know about AWS Activate? That's their program for startups where they provide up to $100,000 in AWS credits for all startups. Whether you're backed by an investor or you're bootstrapping money is time to keep innovating, time to delight customers and time for you to keep gaining traction. You need Runway and Activate is going to help you with that Runway. We hear this story from so many of our founding university companies are finding product market fit, the words getting out about their product. Just a little boost, finding some savings here or there to get a little extra time can be the difference between getting traction and bringing in revenue. Or hey, let's call it what it is, running out of money, okay? And shutting down. AWB knows this. That's why they've created the ultimate toolkit for early stage startups looking to boost growth. With AWS Activate, you're going to get up to $100,000 in AWS credits, Hands on support and training, plus exclusive discounts with some of our favorite companies and tools. So start getting the support you need at every stage of your startup journey. To learn more, visit aws.Amazon.comstartups/credits. That's right. Aws.Amazon.com startups credits. Let's keep moving. I know there's a lot going on in California and. Yeah, here we go. More censorship, possibly.

Yeah. All right, so there's a bill called SB771 over in California. It's passed their version of Congress and is now up for signature by Gavin Newsom, the governor. And what this does, Jason, is create financial risk or a penalty for social media companies that use their algorithms to promote content that runs afoul of certain civil rights rules. It basically sets on precedent that SCOTUS, the Supreme Court, set in the court case Moody vs NetChoice, and that held that social media platforms can create, quote, expressive products when they create a personalized feed for a user. So if you think about section 230, it protects platforms online from the content people post on them. That's very reasonable. I think we all like that. But what this says is if a company has an algorithm that pushes something that does go against certain rules and laws, they can be held liable for it. The penalties are large, up to a million dollars for willful violations. And Jason, if I understand it correctly, those penalties double if the person who is harmed in this case is a minor. Now, we've talked a lot on the show about protecting kids from social media, though it does seem from your tone that you think this is a little bit over the line.

It feels like this is overreaching. I think there's a middle ground that I talked about on previous episodes and on all in just recently and for years on all as well, which is if you give consumers choice with their algorithm and they can byoa bring your own algorithm, then you shouldn't have to deal with this. If you build a singular algorithm that's a black box like TikTok, you should have some liability. You should lose your 230 protection. So if you're TikTok and you were to take a video and your algorithm trended it and that was calling on violence, let's say, and all of a sudden a BLM riot or a January 6th occurs, and you can correlate it to that video, because the people who saw that video then went and charged the Capitol or burned down buildings during blm, let's just say like a violent BLM protest, not the peaceful ones. And if you were to send people to a violent riot and promote violence. So let's say the video itself said, let's break all the windows of the Capitol. And in this mall, in the case of these kind of actions, yeah, I think if you only have one algorithm, you should Be responsible because you made an editorial choice to put that algorithm in. And giving people a free pass on these algorithms is going to lead to the algorithms creating more and more psychosis. This is all science fiction. It doesn't mean I believe in silencing freedom of speech. I believe the algorithms are too powerful. They are making editorial choices, and when you make editorial choices, you need to own them. But if you give people a choice of different algorithms and you subscribe to an algorithm store and you disclose what the algorithm's doing, that would be enough for me to say, you don't need this.

I would also not.

So the.

Oh, I was just gonna say this isn't purely theoretical. Like, there are studies that show, like, there was this one, I was just looking it up in 2021 where there was that study that Instagram points younger girls towards the kinds of content that gives them body issues and anorexia and, like, makes them more concerned about their size. And so we do have evidence that some of these algorithms are pointing people in perditious ways. It's not just purely an academic question at this point. We can look at some of these algorithms. Say YouTube does tend to point people towards, you know, politically controversial or divisive issues, or Instagram does push girls towards, you know, diet content and makeup and beauty content and all of those kinds of things.

I think also, probably part of that is because they're being pushed towards things they can sell ads against. So if you tell people that they're not attractive enough, they're probably going to click on the Sephora ad.

Oh, of course.

I'm not saying it's purely malicious. Like they're trying to give girls anorexia. It's just that's a side effect of they're making the platform as profitable and revenue generating as possible.

Oh, I was trying to make that in reverse, which is saying that the profit motive leads companies to creating algorithms that do harm people. Not that they're creating algorithms on purpose that are bad. They're just following the profit motive. Which is why I think Jason's idea kind of helps people not fall afoul of themselves by giving people more choice. But Jason, lacking the third option that you're describing, your BYOA model, would you support SB771 as it exists or stay where we are today without it?

Yeah, I wouldn't support it as it exists. And I'm just becoming aware of this this week because this is. It just started trending on X and other places and people are talking about it. So as it stands right now, I would rather see something more refined, which is if you have algorithmic transparency, you know what the algorithm serving you, and you have algorithmic choice or no choice at all, you can turn it off. Then I would like to see them remove the liability and have freedom of speech. It's really important because now we're going to be asking this question, well, what should the algorithm do with a really charged instance, like Russia invading Ukraine or, you know, October 7th or Gaza? Like the opinion on Gaza is a moving target. How is an algorithm going to parse that if there's so much misinformation, disinformation, a lack of information? We're asking a little too much of YouTube and Twitter and Facebook to be able to adjudicate that, understand what these algorithms are trained to do at their core and is increase engagement. That's it. They want you to stay on the website longer. That is their North Star. And the algorithms for advertising want you to spend money and click on ads. That's the dual mandate there. So if you understand the mandate of these, then you can kind of understand that. I don't think you could get the companies to tweak these in a way that it would be able to make sense of the conflict in Gaza. I don't think the algorithm can actually make sense of that because humans can't make sense of it and find consensus.

Yeah, it's actually a really good point that you bring that up, Jason, because we were looking through which groups are in favor of SB771 and which groups are opposed. Now, it's not going to shock anybody that listens to the show that technology lobbying groups are opposed to it because it creates the potential for financial penalties against their client companies, you know, TechNet, et cetera, opposed to it. Not a shock. But what's interesting was there was a bit of a split between different advocacy groups about the bill. A great number of Jewish community organizations, for example, are in favor of SB771 because they view it as a very particularly strong tool to combat antisemitism online. And we have seen data from the last couple of years that antisemitism in the digital realm has increased in frequency and quantity and badness. So that's, that's a fair point. On the other side, Jason, back to your point about not being able to parse very complicated issues. Many Arab and Muslim advocacy groups here in the states are opposed to it because they think that it's going to lead to platforms censoring, perhaps pro Palestinian commentary. So it's a bit tricky to decide what to do here, though I do think I fall on your side of the fence here. But, Lon, I want to get you in now.

Yeah, I think that's a real. That really highlights the warning and the danger of this kind of legislation. I believe it's well intentioned. I think that, like Democrats at Gavin Newsom are trying to do something good that's trying to arm groups like, you know, LGBT advocacy groups to help them, you know, combat some of this hate speech online. I think that's fine. But the downside is, as you've seen, well, Zionist groups are saying, you know, if we define our position as if you're opposed to it, that's antisemitism, as we've seen some Zionist groups do, then they can use this as a cudgel to suppress pro Palestinian speech online. And I think that's why you're seeing this divide, that Zionist organizations are saying, oh, this is a good thing. We can help police more speech online. And the pro Palestinian organizations are recognizing they're the ones that are about to be policed. And so I feel like that does show the danger is like, well, if you. Who. Who are you entrusting to decide what is misinformation, what is hate speech, what crosses the line? I think it's too ambiguous for a law to designate.

Hey, listen, we meet a lot of early stage founders here at Launch, my investment company, and some they don't have a lot of traction yet. They just have an idea. Maybe they haven't even finished their product, they've just got an mvp. But they still need investors and accelerators like ours to take them seriously. And you know what? We can't just wire money to your gmail or your PayPal. That's not how it works, folks. We need to know that you're a legit and official business. We need to know your company is incorporated. That's why you need Northwest Registered agent. It's the service that will help you run your business the right way from day one. In 10 clicks and in under 10 minutes, you're gonna file for your LLC or a C corp. If you're a startup, get a domain name, launch your official website, claim your business email, and even fast track your trademark application, which some people forget to. We're talking about more than just company formation. This is your entire identity as a business. Go to northwestregisteredagent.com twist and show the world you're in business. And make sure you use that URL twist so they know that we sent you, I will add there. So. Very well said. The other issue, I think, or the opportunity will be if you hold these corporations to an unreasonable standard, then people will move to distributed systems powered by crypto. And the reason people haven't moved to those systems is because the current systems are working just fine. But if you did start to create issues with the current systems, I think Blue sky and different federated services would become the outlet that people would move to, because then you can't be censored. Right. And so if I'm putting stuff onto a blockchain and then you're using distributed computing to pull your feed and build it together, I think that's where this is going to wind up. Now. Then you're going to have a dark web effect where people are going to do sinister things on distributed computing, like publish, I don't know, thousands of spam accounts in coordination using AI. So, you know, pick your poison, folks. In a distributed system that nobody has ownership of, it's up to the users and the nodes to try to make sense of it. And then with centralized, you know, it's up to the corporations to do this. I think it's unfair and I think it relates to the YouTube story we're about to get to.

Yeah, beautiful segue, Alex. No, go ahead.

We don't know if Gavin Newsom's gonna sign this bill, Jason. I have a suspicion he won't. And then he'll wrap himself in free speech garments afterwards and then run for president. So this is not a done deal, everybody. We're still debating this. But, lon, back to you.

Oh, I was just going to say this leads beautifully into our next story about YouTube and how they had attempted to sort of police speech on their platform. So in a letter to the House Judiciary Committee and its chair, Ohio representative Jim Jordan, YouTube's parent company, Alphabet said it's going to reinstate the creators who were banned for spreading misinformation. I'm doing air quotes if you're listening. On Spotify. They're banned for misinformation about COVID and the the last two political elections. Here's a quote from the letter itself. YouTube values conservative voices on its platform and recognizes that these creators have extensive reach and play an important role in civic discourse. The company recognizes these creators are among those shaping today's online consumption. Landing must watch interviews, giving viewers the chance to hear directly from politicians, celebrities, business leaders, and more, unquote. Alphabet also implies that they were directly pressured into action by the Biden administration. Another quote. Senior Biden administration officials, including White House officials, conducted repeated and sustained outreach to Alphabet and pressed the company regarding certain user generated content related to the COVID 19 pandemic. And here's the key part that did not violate its policy. So they're saying people were posting things that were not against YouTube's terms of service, but then the Biden administration was saying, you need to delete that stuff anyway, and then they complied. So that's. We are taking another step down that road of, of, of looking back on the Biden years and saying they there legitimately is an argument that they were suppressing speech on some of these platforms.

Yeah, they obviously were both of these parties. I hate to break it to folks who are into these two parties. I'm an independent. I'm not part of either one for a reason. And I want a third party as soon as possible, please. That's where I'm putting my support going forward. After my delving into politics over the last couple of years, being dragged into it by my friends because Foushee himself was figuring out ways. And the team over there with the FOIA lady, remember all that, they were figuring out ways to use hackerspeak, trying to figure out how to delete emails because they were having the same conversations that the quote unquote conspiracy theorists were having. And Jon Stewart pointed out with his famous bit on the Wuhan lab having a Covid Wuhan lab, you know, so I think a lot of this is great because we can kind of learn from it as a society that censoring people is wrong and you can sense, I mean, there's obviously illegal speech, but what is the government doing trying to. If somebody is wrong about medical information and they're posting their crazy theories about what they think, you know, meditation or, you know, oils, essential oils, or, you know, their crazy diet. Like if you have a crazy diet and you're doing all that stuff, like, okay, it's up to consumers to make that decision. If I'm going to listen to the crazy yoga lady who says just eat fruit all day long or I'm gonna listen to Tim Ferriss who says to go into ketosis, like, it's not the role of the government to sort all that out in my mind, yeah.

The thing that I'm curious about is whom are we going to grant a little bit of power to? Because I'm a BIG advocate for Section 230 protections for platforms to be able to set their own rules. But there are people who are also saying that Section 230 protections should be stripped away, which would lead either to a ton of censorship online or an absolute mess of Nazi AI slop, as we tend to see. I do also want to say to my friends out there on the left, we're talking about the Biden administration doing what's called jawboning, which is essentially getting people on the phone and hollering at them. Not using a law to force action, but essentially browbeating them, working the refs, if you will. I mean, personally, I don't think it.

To be fair and even handed. And look, I'm no Trump administration fan, but that is basically what people were also accusing Brendan Carr of doing last week. There was not a law passed that Jimmy Kimmel had to be pulled from the airwaves. He just went on a podcast and looked into the camera and said, we're going to do this the easy way and the hard way. It was. It was soft power, just like the Biden administration was sort of doing.

I would say threats of license revocation is a bit different than job voting, but I'm trying to say, lon, to your point, is that we shouldn't give people on the left a pass on this issue. This was a bad thing. Job owning is, in fact, bad. I just want to say that platforms, then, if we're going to take the government out of this entirely, therefore become the ultimate arbiters of what is allowed and is not allowed. And therefore, to Jason's point, consumers have to vote with their feet, or, I guess their eyes, in this case, in their ears. But I've been in completely uncensored online spaces before in terms of there haven't been no rules, and they don't tend to work that well. So what we're going to do here is let corporations and platforms set the.

Terms, which they should, and then you can choose which one you want if you are. You know there's no adult content on YouTube, right? Like, they draw the line at pornography. There are platforms like LinkedIn and Facebook that require real names. There are some that require pseudonyms, so you can post to it, but you have to use the same handle each time you post, like X and Reddit.

And we're all familiar with, like, the rumbles of the world. There are parlors. There are all kinds of, like, alternate X if you want more conservative viewpoints or you want to be sort of less censored in that way. It's a common thing.

Rumble, though, does have a terms of service that has an explicit list of things you can't do. I've read It.

Right. I'm sure that they don't allow adults, they don't allow pornography, for example. But I will say that, you know, I'm with you, Lon. I think if you're making threats in public or private, what we learned from the Twitter files, what we learned from this is the Democrats were doing it privately. They were having the FBI go and visit them. Like, the FBI shows up at your doorstep, as Zuckerberg said. Like, we felt threatened, you know, and I would feel threatened if they were like, hey, we need you to take this stuff down. I would take that as the same threat as the FCC chair. It's like FBI officers showing up arguably, to me might be scarier than, you know, and I don't agree with either of them, obviously.

Yeah, I think we're all in agreement here. I just think it's good that we're having the conversation. It's almost good that it involves everyone across both sides of the aisle here in American politics, because we can look at it and I think and all hold hands and say, let's not do that anymore. And that's a very positive place for us to be, to make it a nonpartisan point, because then we can all agree. And that brings joy to my little free speech heart.

I was just going to one thing to add, I think Alphabet is they now said they're going to adopt basically the exact same response we've seen from X. They're not going to officially fact check every video. You know, they're done fact checking. They're going to do Community Notes. So if you watch a YouTube video and you feel like that is full of misinformation now, you will be able to tag it and say, here's why I think it's wrong. And then if enough people say that or voted up, you'll get Community noted, just like you do if you're JD Vance and you tweet something.

Can I. Can I throw in?

I got a little inside story here, too, on Community Notes. I was with Elon when he took over Twitter, and there was a bunch of stuff that projects that were going to get cut. And I think Community Notes was one of the ones that was going to get cut amongst many different side projects or whatever. And I said to Elon, like, I think this thing is super promising. You should really double click on it because it actually works. And the guy who created it pitched him, and I watched the pitch. I watched them talk. Elon got it immediately and was like, oh, this seems like the most fair way to do this. You get to have a debate, you get to see multiple ones of these. And in practice, I would say 9 out of 10 times. I agree with the community notes and I have an account that I don't know if you gentlemen have this as well, but I get asked to rate community notes all the time.

Yeah, I think if you have the blue check you do get, because I also get asked to, like, help us out or approve this community note or. What do you think of this? Like, they do seek my feedback. Yes.

What's your general experience with it in terms of. If you looked at 100 of them, how many would you agree with?

Oh, it's nine out of 10, I think. And it's also, I think, impressive how fast they are. Like, a lot of the time I'll see a tweet that's only two or three hours old and it's already got a very lively community note debate back and forth. And like, that's really reassuring because that, I think was the big fear that I had about community notes as like our main attack against misinformation. My fear was, yeah, but they won't get corrected until the next day. So the whole first day you've got the lie going out there and then it only gets corrected. That's like what happens with newspapers. Like, if the New York Times messes up, it's on the front page, but then the correction is like page 23, section C. And so nobody ever sees that. But I don't think that actually happens.

All right, let's keep going. We got a lot to get through here. And.

Yeah, so recall that MIT study from last month where it said 95% of AI pilot programs and experience experiments at enterprises weren't producing, like, real results. And we were all debating like, oh, is AI like, not as good as we thought? Well, now there are research teams from Harvard and Stanford believe they may have the answer. They found that while many workers are using AI effectively to improve their efficiency, make their work better, others are lazily using these tools to create what they call low effort, passable work that ends up making more work for their colleagues down the road who have to go back, adjust it, fix it, make it better. So they define workslap as, quote, AI generated work content that masquerades as good work but lacks the substance to meaningfully advance a given task. So a lot of the. A lot of this is based on a survey. They only sampled 1,150 people, which is a lot, but not massive. So keep that in mind. Grain of salt. But 40% of the workers they surveyed said they'd interacted with work slop in the last month. And among those who said yes, they had interacted with Workslop, they said around 15% of the total work they see these days qualifies. So I'm curious what you guys think about this.

This is a real thing. I had somebody, I have like three instances of this recently. Somebody was writing show notes for us, and it was workslop or like they were doing recaps or tweets using.

We do a thing called partner clips, where we post tweets thanking the clips from the show that we also thank our partners. And yes, the team. Nobody working here now. We're all, nobody's doing this. But the team was using AI to generate the descriptions for those videos, which would sometimes come out crazy or not describing what was in the video.

And I caught it and I was like, what is this? Who wrote this? And they were like, hey, I wrote it. And I was like, oh. And then I had an analyst who joined a researcher and I had them read out, you know, because I checked these things. And we were like talking about a company. I said, read it out loud, your summary of the company. And he was reading it. It made no sense. This is maybe a year and a half ago. And I was like, wait, did you. You wrote this? And he's like, no, I used AI. And so today, Wednesdays we have our management meeting. So I have a dinner, I bring dinner in and we do a management meeting every Wednesday where the heads come to me and just tell me what they're working on, what their goals are. It's, you know, nice little way for us to get out of our day to day. And I said, I realized everybody was using AI notes then not reading them. And I was like, oh, you know, note taking is very important because when you write things down on a piece of paper, you remember them and you process them. And then if you retype them in or you share them with another person and you crystallize them a second time, now you're really integrating them. And what I found was people in meetings are just passively, like, I don't know, like, not paying attention. So for tonight's meeting, I said, everybody needs to have a pen and paper, laptops closed. I want you to write notes of when you're hearing your peers go over each of their segments. Only a 90 minute meeting. So just write notes. And at the end, we're gonna do an exercise where just for 10 minutes, people type in from their notes, the five most important takeaways they have, and then share them in the group chat. So I'm just trying to get people back to understanding writing. And I think what they're gonna have to do eventually is force young people. Cause I'm finding young people can't write, they can't type. It's like two things they just don't know how to do. This is a serious crisis. And they can't do math as well. And so what they need to do is, yes, they can.

That's everything.

Okay, so kids are coming into the workforce without the ability to do math. I do this, I say, tell me the percentage. They can't do percentages. I ask them to add two numbers together. They literally can't do math in their brains. And these are college educated people from colleges that you would be like, what? So what they need to do is they need to sit kids down with a pen and paper and they need to write their essays with a pen and paper. Yes, I wanna see these kids raw dog their essays. No more using a computer. And you know what? I'm gonna keep going on my rants about these kids being overmedicated. Not that I think this medication doesn't have a place, but it's very clear from all the studies here in America, adults and kids are over medicated. There's too many people who have diseases, quote unquote, that are benchmark through like surveys. And then I talk to these kids, they now have techniques where they know how to get these drugs, they know what to say, and then they additionally know what to tell their teachers in order to get accommodations. And the accommodations are I can take a test with unlimited time with a computer and with the book, which is not a test. That's the opposite of a test.

That's just homework. That's just a homework assignment. But yeah, I think we've lost our ability not to get too down a rabbit hole here, but we've lost our ability to distinguish between a severe case of whatever and then just like somebody who has like one or two symptoms. Like there are people who have severe adhd, severe depression, severe autism, whatever, they need those medications. But then there's a lot of casuals I think that are sneaking through and getting, getting a buzz on that don't.

Maybe really need it bringing us back to work slop. I think this is actually a very important thing to bring up because if you recall the M law and name check at the start of this segment, it didn't say that only 5% of people are using AI at work. Actually, it said that most people are. And so I think what we're seeing here is the same impact we saw in the education world show up in the corporate world. If there's a tool that does most of your work for you, air quotes, people are going to use it. And I think this is going to show off Jason, who is actually useless at the office and who is not. Because if all you're doing is taking your work, putting it into an AI chat interface and then copying and pasting the output, but I can replace you with a $20 a month ChatGPT subscription. I don't need you. So to me, like, you have the option now to turn in crappier work that looks like work, but I think also you're just arguing yourself out of a job.

I do think it's a morale thing too. That's the last thing I would point out here. Like it is a problem in terms of, like, companies are spending this money on AI tools and then not getting good use out of it. It's making more work, but it's also making people resent their coworkers and bringing down the community aspects. There is a chart, I don't know if you guys have this, that they included. 34% of respondents who said they'd received work slop, said they reported it to other teammates and managers today are creating tension within the office. And then 32% said they were less likely to want to work with the work slop creator in the future. And you see here, when people get work slop in from their colleagues, it makes them feel like those people are less creative, less capable, less reliable, less trustworthy, less intelligent. This is the opposite of what you want to see going on in your team. You want people who respect and love working with one another, who are excited to collaborate, not people who think each other is stupid and are waiting for them to send them AI slop.

Yeah, Workslop is real. This is just people are gonna think that you're not intelligent and trustworthy. That's not a way to build a career. Hey, let's talk a little bit about some prediction markets. We saw that south park is going to have a whole poly market, I guess, episode. But I want to understand what's going on with the TikTok odds. So walk us through this, Alex. Yeah.

So as we've talked about on the show, there is the rumblings, the makings of a big TikTok transaction agreement between both China, the United States, ByteDance, the Chinese company, and also A lot of American interests, both financial and technologic. And so there's an interesting market over on Polymarket trying to sort out when the TikTok sale will be announced. Now, Jason, what we like here is that it gives us a couple of different options. Three, in fact. Will TikTok sale be announced by September 30, October 31, or December 31? And as you can see, if you're on the audio version, 40% chance this month, 70% chance roughly by the end of October, and roughly an 80% chance by the end of December. So people are really betting this is going to happen, but not in the next couple of weeks, probably in the October timeframe. And what I love about this market, Jason, is that there's $1.4 million in betting volume. So we have a lot of folks raising their hand. Now the question is, who's right? Jason, lay your bet.

Well, you know, I always tell folks to look at the details and the rules of this Poly market. This market will resolve to yes, if ByteDance ByteDance announces their intention to sell TikTok by September 30, 2025. Otherwise, this market will resolve to no. A public announcement of the intention to sell TikTok will qualify even if a sale has not been finalized. This includes any formal declaration made by ByteDance for its authorized representatives regarding their plans to sell or transfer ownership of TikTok. The primary resolution source for this market will be official information from ByteDance. However, consensus of credit reporting may also be used. So it has to be ByteDance saying it, and the source has to be ByteDance or other people reporting that ByteDance said it, but they don't have to close the transaction. So this is what's critically important when you're looking at these prediction markets, including my favorite, polymarket. So when we look at polymarket here, it's has TikTok announced this sale? Not that they've closed it. Right. It's just the announcement that they intend to do it, which means it might not happen. I feel like 80. 80% chance by December 30th is a lock. I feel like that's a lock. I think that might be found money.

I would actually say December 16th, because that's the Trump extended the deadline to December 16th. So if this deal doesn't go through by December 16th, Trump either has to once again extend the deadline or TikTok has to cease operations in the US so I bet there is a plan and they're planning to resolve it within the first half of December.

Not, I now think that it's a higher chance that it'll be after December 16th. Because if you've seen one thing in this TikTok saga, it's that the law doesn't really matter here and we can just keep kicking that can down the road. So why not let's extend it to 20, 28.

I mean, you know, you should buy the no. If you put $10,000 on the no, you win 21,250 bucks.

I would make a difference between me joking at LON and also putting a small used car onto polymarket.

Oh, no, wait, wait, I take it back. Yeah, that didn't make sense. It should be roughly 4x. So $1,000 gets you 3,800 basically. So I might, I like the odds of the no, actually, because what if things break down with our negotiations? Polymarket is going to be available in the US and I think we, since they're partners, do we have like a, a promo code for them or we just say polymarket?

I guess we just tell people to go check out their check out polymarket. I don't believe we have a specific Polymarket call to action, but go, go look at polymarket and if you're outside the US go, go buy, go check out a market. Go, go lay your, lay your bets.

Okay, let's go to the Tether story. Alex, I'm sure you're excited about this one.

Oh, absolutely. Hype. So the news everybody, is that Tether, the company behind the well known USDT ST, which is pegged to the US dollar, is in talks to raise between 15 and 20 billion dollars. Now the company, which is based in El Salvador, could be valued at up to $500 billion in this transaction, the old half trillion dollar mark, which as we all know is roughly what OpenAI is valued at today. So that means that we're looking at about a 3% stake for the money here. Jason, this strikes me as a relatively aggressive fundraise for a profitable company and a valuation that makes me a little bit nervous. But your first reaction is.

Well, you know, they're making 5% or so on Treasuries, I guess something in that range, 4 or 5%, I guess, depending on the timeframe that they're holding those. The company has had a checkered, sorted, controversial past I think is a fine way to say it. They've been banned in certain markets. There's reports that they were insolvent potentially. They didn't do audits and so they have a certain period of time to clean all this up. I think they're well within that. They're being advised by Cantor Fitzgerald and I think Lutnick was from Cantor Fitzgerald, although he's not anymore. So I think if you were to look at the stablecoin alignment, all of this has been a big setup for Tether to go legit. And now that they're profitable, because they have so many tethers out there, it's well over $100 billion. I understand somebody can check producer Claude for that, that if they're making 5% on that, I think they started releasing their actual revenue and it was tens of billions of dollars a year. So this is extremely popular. However, rate cuts are coming. We just had one. As rate cuts happen, they are going to see their business collapse because they will not be able to make the float. Also in the Genius act is that they're not allowed to give interest to people who own stablecoins. So people who own stablecoins can't get. If you own a lot of them, you're not gonna be able to get interest on it. I think what this will create is places like Robinhood, which I'm a shareholder of, or Coinbase are gonna have their own stablecoins. But what they'll do is they'll do automatic sweeping. So in most sophisticated consumer first products, you have legitimate ones, the ability to automatically sweep from your non interest bearing account, like checking to your interest bearing account. Most banks will not do that for poor people and unsophisticated, quote unquote unsophisticated people. But banks for rich people do this automatically and I think Robinhood does the sweeping automatically. I believe that this business will collapse. The interest on these businesses is going to collapse. So they are selling at the peak or they should have been doing this six months ago when they actually had a window. But I guess because the Genius act happened, you know, they did this. But Coinbase has a product with Circle and they do give some amount of interest. Correct.

Alex, maybe you could explain 4.1 and 4.5%. But wait, didn't Jason just say that the Genius act bans stablecoin issuers from paying interest to stablecoin holders? Absolutely. But Circle and Coinbase have long had an agreement that Circle gives to Coinbase the money that it earns off the reserves for USDC tokens. It's stablecoin that are held on Coinbase's platform. And because Coinbase doesn't issue them, it can essentially remit all that money to the people who hold the stablecoin on its platform. So what we have here. What a loophole. And there's been some controversy between Coinbase's legal team and the banking industry about if this is fair. But I really think your point about a collapse is very important because why would I want to go buy a tether stablecoin in the US when I could hold a USDC stablecoin also worth $1 and earn interest against it? So Coinbase has found a way to to make the stablecoin market less profitable for the stablecoin issuers. Not great for Circle long term, but I think it really just goes to show that the amount of money that Tether is currently making, and according to their own attestation, Jason, they had $4.9 billion in profit in the second quarter, is going to be diminished, not just from the Fed, but also through competition. Everyone wants to have a stable.

Why wouldn't I long stablecoins? Because I think they're going to 20x from here. I think they're going to replace things like PayPal, Venmo checking accounts. So if we all go to a concert in the future, I think, or if we all played poker together, we will be settling up with Circle or Tethers rather than using Venmo, PayPal. And those things, why we're going to have wallets. Those wallets will not be tracked. There will be, you know, less and there'll be less fees. So I don't know what the fees are on the Venmos of the world, but they're in PayPal's but they're incredibly low. And I guess on some of them it's free to even send between family members or incredibly de minimis. But this is all going to be massively deflationary. So it's great for consumers. And the real losers here are not the Stablecoin companies. They're going to be ultimately winners. It's going to be the credit card companies who are going to lose. Because don't be surprised if you go buy some furniture and the person says, yeah, that's a $2,000 couch, but I'll give you $50 off. I'll give it to you for 1950 or I'll throw in these pillows if you stablecoin me.

So they're the new cash then, Jason, because you can often get a discount for using cash as opposed to a credit card.

This is the new cash. These will be considered like cash because it won't have the typical 3 or 4, 2, 3, 4% merchant fees. Merchant fees are going away. I would be short the visas and mastercards and Long, the stablecoin companies.

I was going to say, I have some answers for you both. Tether is currently holding about 127 billion in U.S. treasury exposure. So that's answer to that one. And then for PayPal, it's about a 3% fee to sell goods or services using PayPal. And then there's also a fee for currency conversions. Those are the only fees for Venmo Instant transfers are a 1.75% fee.

That's not nothing.

Not nothing.

One last data point here, Jason. I was really curious about the comparative valuation of Tether at 500 billion and Circle, which is public at about 33. Yeah, $33 billion. And using their Q2 numbers from both companies, Tether's worth about $3.18 in equity value at 500 billion per tether in circulation, where a Circle is only worth about $0.44 per dollar of its stablecoin in circulation. So we're seeing a rather large valuation gap there. And I think that's the margin we're discussing is a bit rich. In the case of Tether, well, here you go.

Private company is worth essentially six times more or seven times more. Seven times more than the publicly traded company. Publicly traded company has a bunch of people voting every couple of microseconds. So now you've got millions of people voting on the valuation of Circle. And Tether's a private valuation. It sounds to me like tether is worth 100 billion, probably double USDC. And so if they do go public again, I think this will be a bad trade. If I would never make the $500 billion, if you offer me to invest in this at 500 billion, I would not do it. I would not put a dollar in. Now, if you asked me to invest in Circle, I haven't made a bet on Circle. I actually would. If you gave me $100 to invest, I would put a hundred of it into Circle and zero into Tether at this valuation. If the valuations were in line at that $0.44 per dollar of stablecoin, and if it was equal, I would just split it equally. I think they both have a chance, actually. I would probably give Tether. Well, I would give Circle more. I give Circle two there. It's Tether one third as a private company because you don't have as much oversight. And I would trust the management of Circle 100 times more than I would trust the management of Tether given their performance. That being said, the fact that the Tether project has been a bit of the Wild West. You could also, as an investor, say, okay, they're willing to do things to make money that maybe Circle's not. So it's kind of like a Zuckerberg conundrum where, like, Zuckerberg will bend some rules to make it work. Anyway, really interesting story, and we are going to keep tracking stablecoins.

Episode 2004 was the last time we had Jeremy Allaire from Circle on the show. If you want to go back and dive into a interview with him, I think I did that one. Great chat. All right, moving along. Jason Stripe is going to purchase shares back from its investors. According to Axios. Now it has a new valuation, a 409, a valuation of $106.7 billion. But this is not an employee tender in which they allow employees to share some of their stock. Sorry, sell some of their stock to other shareholders. Instead, the company itself wants to repurchase equity from its venture backers. There's very little precedent for this, but I think it as a way for the company to provide a little DPI and also to reduce pressure to go public. Jason. But I can't recall other examples of this in the past. Is this de novo?

I think this has happened in the past, but not in the circumstance of an incredibly successful company. Sometimes companies that decide they're going to be a profitable enterprise run it for profits, not go for growth where they're trying to grow 50% year over year or 30% year over year. You could have the venture capitalists ask to sell their positions back. We've had to do that in a couple instances as seed investors. Sometimes a founder builds a business that makes a couple of million dollars in profits, and they're just like, you know what? I kind of just want this for my life, and I respect them. But then they have to say, okay, you invested when it was a $5 million company, it's making 5 million in profits, which times 20 makes it $100 million. So what are your shares worth? And then that's where you have this weird negotiation because there's not a market for those shares because the founder has said, I'm not growing this, to take it public, et cetera. I think this takes a lot of pressure off them. If they are profitable and they have money sitting around doing a share buyback as a company in year 12, 13, 14, whatever they're at now, sure. Then they don't have to have those venture capitalists have protective provisions, board seats, et cetera, and they don't need to go to them with their information or get sign off when they want to do transactions. So, so this is a super power play. If this is a wildly profitable company, they should just keep offering their shareholders and then get rid of those shares. Eventually then the founders and the company will own the entire company. Uber's buying back 20 billion. There's gonna be 9% less Uber in the world. When they announced that a bunch of different institutional holders bought the shares and that's why it broke a hundred, because they're like, okay, this business isn't going anywhere. There's going to be less shares available, therefore the shares should be 9% more valuable. And it went from 90 to 100 and it was like, okay, it's now 10% more valuable. And then people are like, oh, this management team understands what they're doing. So this makes you de facto more interested. But these are weird cats. Like they're unique cats. I don't want to say weird. They're unique cats in all the world. The Collison brothers, they have said they might see themselves never taking this company public and have a 30, 40, 50 year old company and never go public. So if I was an investor I would sell half my shares at 100 billion. Sure, you got to get that DPI. In fact, Sequoia did this where they said, hey, you have the opportunity as an early investor to sell shares and we're going to be letting the late stage funds buy those shares. So this is in private equity you have this ability to swap out one group of investors for another. I think this is going to become more common but not the default case.

Roelof, actually Roloff both have talked about this at the all in Summit, how Sequoia is like they're sort of holding on to some of the equity even after these companies they've invested in go public. And just it's a very different kind of business for them. Just kind of like sitting on these stock profiles long term from these companies that they help to seed and grow. So that's interesting.

I mean I own shares in, I own every share of Robinhood I ever bought as a private company, which was probably at pennies. I own a large amount of my Uber shares and again in pennies and then new bank Door Dash and a bunch of other companies that I was invested in as a private investor. I still own those shares because if you believe in the management then and they're still cooking with oil, why would you sell? The only mistake I made was I distributed those public shares of Robinhood to fund one at $12 or something now 120. So. But you know, my investors understand that. So I just had one of my invest two of my investors in that first fund who had put in like 50k. I think the fund might be 5, 6, 7x if you didn't sell your Robinhood shares, not counting the comm shares, the superhuman shares and other liquidations we've had, that fund will probably wind up being, gosh, I don't know, it could be a 10 to 15x fund when all said and done over 15 years, not 10. And so there is something happening here and it built a lot of credibility because I had said to the people when I distributed the Robinhood, I am personally not selling my Robinhood shares. I believe in the company, I believe it's undervalued. And I in fact bought more shares at $10 when I was doing J trading, which we got to get the jtrading.com back going here. And I got to do a partnership with one of these platforms, maybe Robinhood, to just do my J trades and let people follow them. Because that J trade. Let me look it up here. Hold on. I gotta do a. I'll give you guys a little J trading insight here.

Doing a live. A live J trade.

Well, I bought Baidu the other day. Oh. Because they're self driving unit.

So I think that that's Apollo Go, right?

Yes. And Baidu. Yeah, I just bought Baidu. That was my most recent trade. But just looking at. I bought 5,000 shares of HUD just out of complete annoyance that I was trading at $10.42 a share. I am up 1,100% in those two years since I bought it. Yeah, I'm up $584,000 on that purchase.

I want to raise more than I.

Declare those shares collectively in my mind.

My Meta shares I bought. Here's my Meta. I bought a bunch of shares. A bunch 502 shares at $97. It was at 94 over the weekend and then I heard him do the layoffs that weekend and I put a market order in. So it went up like three bucks, about 677% on that trade. $332,000.

There were a few years there where you could wait for Zuckerberg to get himself in trouble or for there to be a public controversy and Meta shares would dip like 10%. And if you bought the dip, it would inevitably go back up and then it would crash again for the day. And like you could just keep playing that like riding that wave over and over again. And make. I was making a few hundred, maybe a thousand bucks every time.

Yeah, you buy the bad news and sell on the good news.

People are going to stop using Instagram and Facebook though. The WhatsApp isn't going away. So yeah, he's got a bad news cycle or two. They'll ride it out.

I think this is my next biggest win. When I did J trading, actually I did it in 2022. So it's been three years. Sorry, three years. Up 95% on this win in three years. My average cost of $112 now trading at 220. Amazon. I bought 2,500 shares. Market value 550 dimes, up 269. I just did this for fun because I had these in like a QQQ or just like a blended or some retirement account and I was like, I'm just going to day trade this. I think it was in my Wealthfront account and I just started day trading, was it. I think Amazon is my big pick. I'm going to buy more Amazon people are down on Amazon. My belief is Amazon's going to shed. Ask producer Claude how many employees Amazon has and then what percentage of them are drivers and factory workers versus executives. I think they have hundreds of. I think they might be close to a million. They are one of the largest employers. I believe the factory workers and the drivers are going to go down 10% a year for like five years in a row. It's going to be unbelievable.

Amazon had 1,546,000 employees as of June 30th of this year. 1.56 million full and part time workers is a good estimate. Approximately 65% of Amazon's workforce, about a million employees work in operations and fulfillment roles. That's warehouse associates, delivery drivers, other logistics coordinators. That includes both warehouse workers and drivers. We don't necessarily break them down individually, but Claude is saying about two thirds are involved in the manual moving around of products to get you the thing that you want ordered.

So if they're growing 20% a year in sales, that means that team, I wonder what they're growing at. We could make a really interesting poly market of how many of those class of workers they'll have. But given what I'm watching with Optimus Figure Robotics, Zoox Robotaxis, Waymo, I saw that Neuro and the Lucid just had their first pilot or they got the first mule and they, they put the sensors on it. So that's moving fast. I think you look at all of that activity, I can't imagine that they will ever add more drivers or I think this is peak drivers, peak factory workers. And I actually had somebody from Amazon cause they heard me go on this rant and I had somebody correct me on Twitter and I had a comms person be like we haven't actually lowered it. We found new things for those people to do like same day delivery, like other stuff. So I will give them credit that they are trying to redeploy folks intelligently to not have those jobs. But that will be the one place where we're gonna see the most acute job destruction. Now it's a million of how many workers we have in the United States? 62% labor participation amongst roughly 280 million Americans. So that means we have 150 million people working. So it's not even 1% but it's a significant group of people who are going to be the first to go or amongst the first to go, which is people sorting through packages and putting them in boxes.

I just want to add, you can't do everything we're doing with producer Claude without a paid plan. But if you go to Claude, AI Twist Twist listeners get 50% off their first three months of Claude Pro. So go check that out.

Very generous. Okay. You had something to add there, Alex?

Oh, I just wanted to. I was going to take us over to a founder question that I wanted to get your take on, Jason.

Love it. Let's end with the founder question. I love it.

All right. So I do spend a lot of time reading the startups Reddit, a lot of great folks over there and this question really took me by surprise, Jason, because I didn't have an answer for it. So Redeemer over on R startups wants to know how do you negotiate pilots so they don't stall out after success? They're running a pilot with a quote multi location client and they're worried that after it goes well, hopefully that they're going to try to renegotiate or it's going to stall out. So if you're an early stage company doing pilots with some of your first customers, Jason, how do you ensure that after the trial period you actually end up with a paying customer at the end?

Yeah, so you should keep. You should have a discussion of hey, what would success of the pilot be for you as a customer? And then shut up and listen to them. So what notes do we need to hit in order for you to have a successful pilot and put us into production? So if you were figure and you had BMW, remember there was this Controversy of, you know, are they an actual customer or not? Figure should say to them, hey, we've sent you these five robots, we've got five engineers, you know, one behind each of them, and we're doing this pilot with you. What notes do we need to hit for you to turn those five into paid or if it is a paid pilot, how do we go from this modest paid pilot to a full blown pilot? So you just have that negotiation in advance, but you can couch it in, what does success look like for you? And then what would the deployment schedule be like? So if we're talking about the Robots, we have five robots here, and I know you have 20 different factories. Which of those factories would you want to go to next and how many robots would you want to put in them and what would the pace be? Would you want to do all 20 factories at once with 10 robots? Would you want to do 10 robots in each factory and then separate them month by month? Would you like us to do four factories at a time with 10 robots each? Tell us what would work for you and how do you look at it? We're not going to hold you to it, but it's helpful with our planning. So you just kind of have that discussion. But at the end of the day, if you make a product that they can't live without and you study their usage of it, then you are in the catbird seat and then you can say, you know, that doesn't work for us. And then having multiple customers and multiple pilots then gives you the ability to say to one, hey, we have to focus on this other customer. So it's been great working with you, but we have limited bandwidth and we need to focus on this customer. I know this because we have a lot of people who want us to bring Foundry University to different cities around the world. And I've said to my team, I think we can add one per year. So the first one will be in Saudi, in Riyadh, the first week of November, it's starting. And the second city, if people aren't serious about it and they don't, you know, the pricing doesn't work for them, the commitment and the time they have to put it into it doesn't work, that's okay. It's our job to look at the five to 10 people who want to do it with us and say these three are the most likely. And then this is the one we're choosing to go with. Right? So. So you also have to be willing to fire your customers or pick your customers and the way you get the right to pick your customers is by having a really great product that's priced, you know, appropriately. And if people are going to nickel and dime you, my Lord, it's never going to end. Especially the big companies. When somebody's nickel and diming you at a big company, and I've had one of the big companies nickel and dime me who, you know, typically sponsor stuff and I just told them at one point, like said, I just, please stop emailing me to collaborate because somebody new takes this job and becomes the evangelist for this company. It's big established company and then they try to do a partnership with us. It's the same thing every time. You want everything for free and you want us to do your job for you, we don't do that. It's a hard pass. So if you can't compensate us for the work we're doing and pay our team, and you're from a company worth billions of dollars or trillions of dollars, whatever it is, with billions in profit, why are we even in a collaboration here? You obviously don't respect us. And I literally told them, stop emailing me.

Yeah, exactly. Jason, a follow up question about this Pilots as a thing that startups might get into, when do pilots make sense? When should they be avoided? Because I don't want people to end up in a never ending stream of doing kind of trial wear for customers.

I mean, a pilot is a way for you to get a paying customer who's not ready to pay for your nascent product on board. If your product was fully baked, it wouldn't be a pilot, you would just be selling them the product. So typically when you sell a pilot into a company, it's because your product's either not completed or, you know, fully completed or it's such an innovative thing that it would disrupt their existing way of doing it and they just want to trial it. But that's why I'm only for paid pilots. No free pilots. When you put a price on it, you get rid you, you then have a commitment from their side, which is critical because if you don't pay for something, you don't actually use it. How many times have has somebody sent you free software that you never used? If somebody sends you a video game, if you're into video games and gives you a free copy, you're not going to play it. But if you paid $99 for it, you're like, I got to get 10 hours of this at least to give it a shot. Right? If you bought tickets to you know, the new Paul Thomas Anderson movie? You're going to go and you're gonna watch it in the theater. Alex, why don't you take us out with a Get Startup Credits ad read? Let's see you do an ad read here since I have to carry all the ad reads. Here we go. Go ahead, Alex. It's at the bottom. Get started. Credits. Yeah, I'm right at the sign off.

Yeah, credits. Ah, there it is.

Yeah, go ahead. Let me see your ad read. Style here. The stylings of Mr. Alex Wilhelm.

I feel like I'm being set up to fail here, but ladies and gentlemen, let's do it. All right, now, we all know that Jason invests in an enormous number of companies. We profile an even greater set here on Twist. And that means that we talk to a lot of founders. We are absolutely plugged into a huge network of businesses and services, and we want to take all that networking, all that time and give it to you in the form of startup credits. That way, your early stage startup can beat the competition and also reap the benefits of hanging out with us here on Twist. So what do you have to do? Just go to getstartupcredits.com, sign up, give us a little basic information about what you're building, and then we will custom match you with tools and resources that are tailored just to what you need. You can get over $100,000 in free startup credits right now. Just go to getstartupcredits.com and feast your eyes and your wallet upon the work we have done for you because we love all founders here at Twist.

Nailed it.

I think. Pretty good. Getstartupcredits.com, i created this because you hear all these promo codes, you forget them. So it has all the promo codes and direct access to all those partners in one form. So you fill out the form once and then you get all these credits, including the Google credits, everybody. And so we're going to keep building this. If you want to be part of the Get Startup Credits program, you can email PartnerLaunch Co and yeah, we'll get you or I think this week in startups.com partners. I think there's a form there to fill out where you can become a partner of the show. And so this becomes like, my idea for making the ads more effective and helping founders, because founders always want, you know, the best deal possible. So we just tell everybody, give us the best deal possible, and we'll put it into one thing. And then they get your contact info. So you don't have to give your phone number. If you wanted to, you could. If you want to use like your email or if you've got an email for marketing stuff, you know you can use that email. But it's just a great way to balance out not cooking people, having them opt in to getting credit. So I really, I'm going to be focused on this in the next year. I'm trying to get like another 20 things in there so founders just get everything at once they need. So if you have a product or service for founders, we're now getting, I think hundreds of founders filling this out every month. And we're just starting to promote it because it's working. Okay, that's been a great show. Thanks a lot. Thanks, Alex. We'll see you all on Friday. This week in startups.com is our website and we go live on the YouTube, on X, on LinkedIn, on Substack. So sign up for all that great stuff and we'll see you next time. Bye bye.