INSIDE How AI Startups hire, AI Roundtable with Wade Foster, Mikey Schulman, and Ali Ansari | E2225

It is true that people are getting isolated and that people aren't leaving their homes. I think it's a double whammy of the addictive nature of this technology and COVID just making people into homebodies. What are your thoughts, Mikey?

Most of this isolation does not come from AI. It predates it. It's a lot of social media. You can say there are pros and cons of social media. Imagine you can give the entire world a median competent therapist with AI. Like, that's probably good. For some of the mental health crises that we face in this country. To me, it's just like somewhat incongruent to say that the correct solution to that problem is to ban AI. You're probably going to make things worse. Human fulfillment is really important. Probably working is an important part of being fulfilled, but it's actually kind of amazing when you see some of the stories and you know how many people you make smile every day.

How are you thinking about raising kids into this crazy addictive soup?

Well, I'm fortunate my kids are young enough that, you know, this is not yet a problem, but I am starting to think about like, you know, the kind kind of environment we want them to be raised in. And I do think, you know, it's important that, that they are involved in the community. And so we want them to be in person at school, we want them to be in dance and gymnastics and sports and music and all this other stuff, because they think they get to interact with other people and they get to learn those skills. We will probably find ways to keep encouraging more of that and limit time with the phone where you're just like on social media and some of these other tools.

This Week in Startups is brought to you by Interpret. Interpret turns feedback noise into customer intelligence so your team knows exactly what to fix and build next. Head to interpret.com/twist to book a demo and see it in action. Squarespace. Turn your idea into a beautiful website. Go to squarespace.com/twist for a free trial. When you're ready to launch, use offer code TWIST to save 10% off your first purchase of a website or domain, and LinkedIn Ads. Start converting your B2B audience into high-quality leads today. Launch your first campaign and get $250 free when you spend at least $250. Go to linkedin.com/thisweekinstartups to claim your credit.

All right, everybody, welcome back to This Week in Startups. And this is a new podcast we're launching in 2026. It's going to be called This Week in AI. Yes, that's right. A new podcast is coming in 2026. We'll be doing This Week in Startups 3 days a week. One day a week, we're going to do our roundtable only with experts, only with builders in the AI space. So think of this as a weekly roundtable in the spirit of what I did at All In or when we do the venture capital roundtable here. And I am stoked for our first roundtable because we found some of the great guests we've had here on the program. And people who are building in AI. So no journalists, no analysts, just people who are actually building. And we've got a full docket today. Wade Foster is here. He's the co-founder and CEO of Zapier, who I have known for a decade. Yeah, huh, Wade?

I think that's right. Uh, we, we confirmed prior that was the last time I was on your show, 9 years ago.

9 years ago, which when you're doing podcasts for 15 years ago, like it's a, you're overdue. Zapier makes you happier as everybody knows, and you started in the integration space. Where if people wanted to, a silly example, but Gary Tan at YC started with this with Posterous back in the day. Hey, you're posting something to Instagram. What if that could automatically go to your Twitter handle? Or you posted something to Twitter, what if that could go into a Google spreadsheet and you would have a record of all your posts? But since then, a little thing called AI happened. Yeah, Wade?

That's right. Got a lot of workflows and agents and all of the above inside of Zapier these days.

And we're gonna get into that in the program because The year of the agent was supposed to be 2025, and we're going to talk about why it didn't happen exactly and what's going to happen in 2026. Also with us from my portfolio, one of my breakout founders, Ali Ansari, is here. He's from MicroOne, which we invested in 3 years ago, Ali, I think.

That's right. Yeah. 3 years ago in the pre-seed, one of the, one of the first investors, one of the first investors.

And I know one of the most cherished, I assume, on the cap table, but the company started, if my memory serves me correct, using machine learning and AI to find the world's greatest developers, yeah? And then something happened during that 3-year journey and you pivoted. Tell everybody about that briefly.

Yeah, exactly. We started as a building an AI screener specifically to vet engineers, and it kind of was born out of a need for my previous company, which was a software development agency. Basically, we were vetting a bunch of engineers, assigning them to projects, and I developed this tool to make that easier for me. And then, uh, yeah, MicroOne at first was this end-to-end recruitment engine. And then we quickly realized that the human data market actually has a bottleneck, which is recruitment, which is, which was, uh, the pivot that we ended up doing.

So now you're helping major language models with their training data. We're going to talk about training data as well. And Mikey Schulman is with us. He makes an incredible product called Suno, which allows you, and we did a demo on this on This Week in Startups a couple of weeks ago. It is absolutely extraordinary that Civilians can now, in the same way Canva enabled, you know, people to make beautiful invites to their birthday party, now Suno is allowing people to make incredible music. Welcome to the program, Mikey.

Thank you. Excited to be here.

Yeah. And we have a lot to talk about as well with training data.

Yeah.

Working with the music industry, which is always fun. So let's start with the talent wars. All three of you are dealing with the fact that we have an investment cycle that is unprecedented in our industry, maybe only comparable to the fiber buildout. And as part of that, not just startups, but also the major companies, the Metas of the world, have been on a tear in terms of hiring people. All three of you have to compete for talent. And we had an insane moment in 2025 in the summer, which seems like, I don't know, 10 years ago at this point, when OpenAI was losing developers to Meta's $250 million offers. And then Just this week, OpenAI decided to end vesting cliffs for new employees, which is a pretty aggressive move. For people who don't know what a vesting cliff is, just briefly, when you join a startup, your first year of equity vests, you basically get it in the, at the end of month 12. Why do we do this? Well, you don't want to give equity to people who don't even make it to year 2. That would seem profoundly unfair. So maybe, Mikey, you could explain to us what it's like on the front lines right now hiring people and what impact Meta over the summer had and now maybe what impact OpenAI's new offer is because this seems like a pretty seismic shift or maybe I'm reading into it. What do you think, Mikey?

I think it's a little too early to know what the impact of the OpenAI thing is. That's like very fresh, but certainly people are talking about it. I think for the Meta thing over the summer, it certainly skews, how everybody thinks about compensation. And even if you're not directly thinking about going to work for Meta, it certainly gets everybody's head in a totally different ballpark for what normal compensation looks like. And so certainly if you don't have that kind of money, you know, we're a small company, we're 120 people, it makes recruiting a little bit harder. For the OpenAI thing for vesting, I just think about it like this. If I'm talking to a candidate who has an offer and trying to convince them to join, and I make an offer with a vesting cliff, and they say, I'll only accept it without a vesting cliff. And I just think, like, what does that say about that candidate? Maybe they don't actually intend to stick around for all that long. I think this worries me just about creating this really mercenary, really transient culture around building companies and something I think certainly smaller companies need to try to resist.

What is the pitch you give them to come work at Suno as opposed to, say, taking a crazy OpenAI or Meta offer? How do you counter that? In the marketplace, practically speaking?

For us, it's a lot about the mission. If you really, really love music, if you really love to work at the nexus of AI and creativity, where there are no right answers, where things are messy but beautiful, I think we're kind of the best place for that. But yeah, like, I'll just say the uncomfortable thing out loud. If that means taking a significant pay cut compared to what Meta or OpenAI is going to pay you, um, it is a less compelling pitch.

Yeah. And so What is the international pipeline looking like? Because we've now, we're in year, I would say, 2 or 3 of this very vibrant competitive landscape. It has to have changed a bit in terms of developers becoming more attuned to being attractive. Ali, I'll hand it to you since you were recruiting developers. Is the pipeline of people who are qualified to work at an AI startup increasing dramatically? Are all developers basically retraining, reframing, getting focused on AI now? And is the, is there hope that this bottleneck will be released or is it going to continue for some time?

Yeah, I think the barrier to entry for developers has reduced by a lot, which in some ways makes, makes it a bit easier. Obviously the supply kind of increases to hire developers, but, but in some ways it makes it actually harder because now you have to really deeply vet folks. Most people can kind of do the take-home exercises pretty easily using AI, and it's, it's hard to distinguish whether they have actually done anything on the exercise themselves or if they've kind of relied entirely on AI. So we, we've doubled down on our vetting process in terms of the developers, and the bottleneck is, is, you know, still there. I think the, the hardest bottleneck is, is, is researchers, like AI researchers in SF specifically. I think, you know, the way we think about it is other roles like you know, finance and operations and so forth, we actually have not had a hugely difficult time hiring for, but specifically AI researchers in SF is like the role that, uh, you know, kind of rejects the offers the most and is hardest to attract, particularly because of the, you know, change in sort of perception and total comps because of the kind of meta postings that have happened. So, you know, we've kind of, we're mitigating a bit by, um, being open to hiring globally. Of course, like in terms of AI researchers, the best talent is in SF, but we are hiring a bit globally. And we're kind of, you know, improving our office space, kind of upgrading the office and so forth to result in, you know, solving that bottleneck.

If you are selling a business-to-business product, you're looking to reach professional customers who are doing business. And the question you have to ask yourself is where do those business leaders spend their time? And the answer is obvious, LinkedIn. And these are the real decision makers. So when you advertise on LinkedIn, you're not wasting your money. You're gonna target your ads based on job title, based on industry, based on skills. So maybe you want a CFO in the technology space who has been in their job for 5 years, right? You want to zero in on those demographics and behaviors. LinkedIn has over a billion members and 10 million of them are C-suite executives. They're going to help you reach exactly the right person with the authority to pay for your product, right? Go right to the top. Use LinkedIn. And we've got a special offer just for TWIST listeners. Launch your first campaign and get $250 free when you spend at least $250. Go to linkedin.com/thisweekinstartups to claim your credit. All right, Wade, you've been running Zapier for how many years? Is this year 12, 13, or 14 for Zapier?

Year 14.

Yeah, 14. And in the beginning, people thought this was like a very niche idea. I remember one of your investors, when you hit like some crazy milestone, I don't know if it was $250 million valuation or $500 million valuation, asked me, should I sell? And I was like, not if Wade's running the company because this is a tool people are going to need for some time to come. But maybe you as the elder statesman here, sort of a twofold question. One, how do you frame the company as an AI company now having started as, you know, a tools company, a SaaS company? I'm not sure how it was originally framed and then How do you, are you losing all of your developers to these crazy offers or how do you keep people engaged in year 14 as a startup? So there's two questions there, take them however you like.

Yeah, so I think the first one is Zapier, in a lot of ways we're really fortunate. Uh, we have 8,000 integrations that connect to all these different tools. And it turns out if you want to build agents, if you want to build agentic workflows, they benefit a lot from having access to data. And so a product like Zapier is maybe easier to adapt in the age of AI compared to maybe some of these other companies that were, you know, sort of built pre-AI. AI era concepts. And so, you know, I think we, a lot of our folks see the opportunity and they feel like it's sort of well positioned. In terms of then like, how do we retain, how do we compete from talent? I think Ali was kind of hitting it on the head. There's like a handful of axes here where I see the like talent battles competing on. There's, you know, AI researchers versus, you know, sort of everyone else. And obviously AI researchers are highly in demand, whereas in, you know, the sort of everywhere else camp, like there still is, you can still find great developers, across the globe. I think you've got SF versus everywhere else. Obviously, the talent wars in SF are very intense. Not everyone wants to live in SF. Uh, you know, J. Cal, you moved to Austin. Uh, so there's a lot of— I moved to Missouri. Like, there's a lot of people spread out across the United States, across the globe that are very talented, and you can tap into that. I think the other axis we see a lot of, um, challenges on are, you know, execs or leaders who are actually using the technology themselves. There's a lot of people who rose through the ranks, you know, in the 2010s who have kind of just like stepped into more of like a purely management role. And those folks tend to run 2015 playbooks and those 2015 playbooks just don't work particularly well right now. And so you have to find, you know, leaders who have like some of that leadership management experience, but are in the trenches playing with the tools, playing with the technology, understanding how it changes the game on the field. And that's another axis that I think is pretty tough to, to, to find right now is people who actually are using this stuff. And so, you know, I think if you kind of look at these different axes and you think about like where you want to recruit from, you got to find the pocket where you have an advantage. You know, for Zapier, one of the things we benefited from the longest time is that we said, hey, we'll hire remotely, we'll hire distributed. And that maybe became less of an advantage during COVID But now that everybody's doing return to office and things like that, it's kind of become a bit more of an advantage again because we're happy to take you wherever you're at.

Tell me about the pipeline of developers. Do we think— and we'll go around the horn here— the Copilots and how easy it is to onboard and become good at UX using these different tools and the onboarding at Replit and just being able to become a developer. Are we seeing an increase, Wade, in the number of developers and number of people who are developer curious? Is that pipeline increasing? Because that was the bottleneck previously. We just didn't have enough people going for computer science degrees and becoming developers. Is that increasing radically because of AI? Less so than you would think.

Yeah, you're seeing a lot of folks who are able to ship and build things. That is definitely coming true. Now, you know, are all of these folks capable of deploying production applications? And that's where I'm like, eh, you know, there is still a hurdle to there where you need to understand, you know, foundational skills to do a good job at building and scaling, uh, tools that you would make commercially available. But even though there is a gap there, it is still a good thing that more people are able to build stuff because it shortens the overall software development lifecycle. You can have you know, a sales rep come in and say, hey, I prototyped this like feature concept that I was talking to this customer about. Like, you know, do you think this idea would be good? And that is like a way tighter communication loop with a product manager or an engineer than in the past where you might have a sales rep come in and just say like, hey, someone asked for blah blah feature. And, you know, it's like, I don't quite understand what you're saying. Like, I can't quite see it in your head. And so it is really beneficial that we have these tools that enables everybody else in a company to get their ideas out of their head and express them in a way that more people can quickly understand, oh, I see what you're getting at here.

And so it's in a way these tools replace the whiteboard session at a company where some product manager or some frontline salesperson is like, hey, we should have this feature. Let me draw for you on the whiteboard. It's like, well, let me just make my own wireframes.

Yeah, I think that's right. I think it replaces the whiteboard session. I think it I think you see a lot more internal tools. I see— I think you see a lot more like throwaway code where it's like, hey, I have a one-off task and I'm just going to do it instead of in the past you might choose not to even do that thing. You just like, I can't do it, whatever, I'm not going to worry about it. And you sort of go along with your day. And so there's a lot more of these like micro problems in organizations that you can actually solve now because you can throw it to Claude Code or something like that to do the work for you.

Like, what are you seeing in terms of the developer pipeline? Is it increasing or is it not a major issue for you because you've intentionally kept the company— I don't want to say small, but, you know, sort of a tight operation. How many total employees do you have, Wade? And how many do you have, Mikey?

We're at 120, but, um, that is not intentionally small. We're trying to grow as quickly as humanly possible.

We're around 750.

750.

Yeah.

So Mikey, tell me, how do you, um, think about the pipeline training and maybe people who are developer curious, being able to digitally with a Vibe coding tool or Claude Code, make these little solutions that maybe aren't production ready, but they certainly around the table make the conversation move faster. Are you seeing that as well inside of Suno?

100%. I'll echo everything that Wade Foster said. Maybe just a fun data point. I think the last line of code I wrote was a little more than 12 months ago, maybe 18 months ago. And then at the time, Maybe 15% of people here were using these things and now literally 100% of developers and many non-developers are, are using these tools. And I think the thing that's super interesting is that when you start to be able to solve these micro problems quickly, when you start to be able to make these prototypes quickly, you shorten a lot of the iteration cycles. It changes how you work. I actually think it makes more of the interpersonal stuff more important. You know, like it, it reminds you how much of a team sport even engineering is, you know, people have some like platonic ideal of an engineer sitting in a room by themselves just hacking away. But actually when a lot of that work happens much, much more quickly, you have to talk a lot more. And I think that's a good thing, right? Like, you know, talking to people is fun, but has totally changed, I think, how we hire, how we develop code, how we develop products. Yeah, it's kind of, kind of remarkable.

It's time to talk about one of my favorite companies in the world, not just a sponsor, but a true friend of this podcast. Of course, I mean the one, the only Squarespace. Squarespace is the number one best way to start a website for your company, and that's going to help you stand out online. Whether you're just getting started with your website or you're upgrading and scaling your existing business like I'm constantly doing, well, Squarespace is going to give you every tool you need: a great domain name, the ability to book meetings, sell content online, and it's going to be gorgeous because they have a new AI-powered Blueprint website builder you can use or you can just use a template. It's up to you. And with a couple of simple prompts with the Blueprint tool, you can design an entirely customizable website that perfectly suits your aesthetic, your voice, your personality. Support TWIST's longest running sponsor by going to squarespace.com/twist for a free trial. And when you're ready to launch, go to squarespace.com/twist for 10% off your first website or domain purchase. Ollie, what are you seeing game on the field in terms of rank and file employees participating in the creation of prototypes? And what are you— more importantly, I guess, what are your customers telling you?

Yeah, I think, I think very similar to Wade and Mikey here, we were, you know, fastening the cycles by a lot. I think in a kind of a paradoxical way, design has actually sort of became the bottleneck instead of engineering. And so we actually, for some of our product buildouts, we're kind of experimenting with not designing first and just having engineers do the, you know, create the prototype and jumping to the iterations that are required on the engineering end. And then once it comes to the state that they believe is functional and, you know, looks good in terms of UI/UX, we then kind of backtrack and create the design files then and kind of fix up some UI aspects afterwards. So this is a, you know, I think a sort of a shift that may happen across organizations where we actually don't get bottlenecked by Figma designs and Figma designs come sort of after the engineer has actually developed the you know, the product. I think the second thing is very similar to what Wade Foster was saying is we're seeing a lot of, you know, admin tools like admin portal tools and kind of internal software being built very quickly in a very lightweight and purpose-built manner where a lot of, you know, software subscriptions become unnecessary. And frankly, the flow of using that very lightweight internal software is actually superior to any subscription that you're gonna make to much heavier weight software. So I think those are two things that we're seeing. And then the last thing I would say is, you know, this results in, as you said earlier, higher supply of engineers because the barrier to entry is higher. So you have to kind of rethink the vetting process. So what we've done is we've moved away from these kind of DSA type of coding problems in our interview process to a much more use AI, build the thing with AI actually, and then just explain your thought process. And then we take that thought process and analyze it and see if there, if they're actually an engineer.

Does this mean we're going to a singularity in terms of startup employees who are valuable, Ali, where the person who can identify a problem, vibe code a solution, cloud code a solution, whatever it is, and then the design is perfunctory and done at the end. It's kind of like we're moving towards one type of person who is the identifier and builder of the solution and just good enough so I don't have to hire some verticalized SaaS product.

Yeah, I think we're going to see a similar trend in a lot of different domains. So I wouldn't say the singularity is like for engineers only. I think engineers operating these machines and became being way better engineers is the same analogy that you can make for recruiters operating an AI recruiter machine and not doing interviews anymore and being much faster at their job and kind of embracing the more creative parts of their job versus continuous interviews back and forth all day. So I think if you apply this same kind of structure of let me be an operator of an AI agent as the, as the main part of my job, and let me try to get to that as fast as possible, then each of those domains will— the singularity of them will be this idea of like running an AI agent, basically.

Wade, what happens to Gen Xers who had like a more verticalized approach where like, I'm a designer, I am a product manager, when you have these I'll assume 30-something Gen Zs, you know, elder millennials or younger millennials even, who can just kind of wrap the whole thing up in a bow and bring it to management.

Well, I, I think the, the skill that is really valuable these days is not how much you know, it's how adaptable you are. And an interesting thing that I see is there's this, there's this debate, like, who benefits most from AI? Is it, you know, the folks who are like senior, the best at their skills, or is it the junior folks who can now work fastest? And I think clearly both are benefiting a lot. But what has been interesting for me to see is the folks who are really sharp at their domain and who are adaptable and interested in learning are just doing some insane stuff right now. And they're able to, you know, collapse the skill stack in a way that, you know, I think was just harder before. Like, it's, it's faster than ever to go from like, you know, 0 to 80 on a particular domain. And so now if you were like a really good engineer who was like interested in design or interested in marketing or interested in some of these other things, and you had good ideas, the types of things you can do now are, are very impressive. And so I, I think, you know, that the Gen Xers who have like, you know, spent their like years in the trenches, like learning a lot, honing their craft and really know how to do that, who are also embracing AI, man, I think you're probably very well positioned to, to build some incredible stuff these days.

Best advice for a young person graduating school who can't get that entry-level job? Wade, what's your best advice to them to break in? Because we are seeing this even in big tech, we're seeing it in traditional companies where, hey, there's somebody on the team who's an elder statesman, as you're talking about. And talking about Gen X versus elder statesman is just crazy for me because we were the punk rock kids, but 15, 20 years ago when Web 2.0 came out. Um, what's the best advice then for a young people entering the workforce who just don't have a job offer from Google or Meta or, you know, Apple today, and they're kind of concerned that nobody is handholding them into a career.

I think like that curiosity is great. Like the thing I remember, I graduated from college in the financial crisis and no one was hiring then, even if you were a good student. And, uh, but this was kind of like the age of the internet. And so like, you know, around when I was graduating, you saw like Stripe and Twilio and Dropbox and Airbnb, like these companies were coming online. And you know, what you could do was you could just go make things, you could build things, you could learn to code, you could figure out how to ship things, you know, and you could email founders and say, hey, I made this cool thing. I think your company could do this, that, or the other. And that sort of like high agency creativity, I don't think that ever goes out of style. Uh, and in fact, I think it stands out even more so today. And so if I was, you know, sort of in that same situation again, I would be using all these tools. I would be experimenting with them. I would be building them. I would be finding products I love, like I'd be hitting Mikey up and saying, I love Suno. This is incredible. I think you all should do this, that, and the other. And here I made a suggested feature. And yeah, I bet good money if I did that, Mikey would email me back and say, hey, let's talk.

Mikey, has that happened? Is that actually happening, that some number of young folks realize sending resumes and doing an AI interview with 100 AI agents and trying to get into a company just not as effective as just doing some spec work, building a prototype and emailing it to you and telling you, hey, Suno doesn't have saxophones that are really great. And I, you know, vibe coded and I got my friend who plays sax to, you know, do a bunch of riffs of Mark Knopfler songs and we fixed it for you. Here's where I think you should be going. Do you actually have people doing that or not? Every day.

Uh, our first Android engineer sent us a little app that he made that was like, we used to have this homepage with this big wheel. Of all the different, like, weird mashups of genres. And it's like, it's on the web and it's like really made for mobile because it's like interactive. And he just built it and sent it to us. And that was our first Android engineer.

You just hired them?

We interviewed him. Uh, but, but yes, then we hired him.

This is just like so obvious, but young people are like, well, if I'm not getting paid for it, why would I do that? The game on the field, Ali, now is you're in a global competition and you, you have your hooks into all the global markets, people in other markets are doing this because they really want the work. Yeah.

Yeah, exactly. I was actually going to say, adding on to Mikey's point, we actually hired our last two AI engineers also in a very similar fashion. One of them built a prototype of our end-to-end kind of interviewer and said, look, I reduced the latency in this one specific part. We should chat. And we interviewed him and that, you know, hired him. Another example was, uh, someone built a very similar model to our proctoring model that exists in the interviews and did a great job and we ended up interviewing him and hiring him as well. So I think, I think I would certainly echo Wade Foster and Mikey's points here, which is you, you, because it's easier to build, you should use that ability and build stuff very quickly and just email the exec team at these companies that you're interested in. And it is very likely that you will get a response back. I think that the other thing I would say is we, We recently actually did a study on the impacts of AI as well as like AI interviews specifically in the, in the hiring market. And what we saw interestingly is that younger folks and less experienced folks are actually benefiting from AI interviews. And the reason for it is simple. You know, if you, if you don't go through an interview process, the only signal you have is one's resume. And of course, if someone is more experienced, the signals in one's resume is by far greater. And so there's this like positive feedback loop that exists that you can't really exit unless you don't only rely on folks' resume. So, so what we saw is that inexperienced folks that are willing to grind and learn and so forth actually did better in interviews and ended up, you know, subsequently getting hired at a higher rate.

A very simple way to say this is do the job that you want. And somebody will hire you. But you just start doing the job every day and somebody will hire you. We took our hiring process and I moved it because I saw Naval did something really interesting. He has got— Naval Ravikant, founder of, uh, co-founder of AngelList, in his feed where he has like this little podcast feed, um, where he just does like 3-minute thought pieces. He just said, I'm looking for a chief of staff and I'm looking for a producer. We are like high-agency independent workers. We work remote. If interested, here's the email address. Tell us what you can do. Show us what you can do. That's it. That was his job description. That was his process. So I was like, well, that's pretty smart. We need a community person here for our founders. And we just popped up the email community@launch.co and said, email us what community experience, what you've done and what you would do for communities around startups and angel investors and seed investors. And we did the same thing. We wanted a CapCut editor. So I said, hey, We want them to do CapCut. Just create CapCut at launch.co and have them email us clips of the existing show. And we paid them for it, but my Lord, the hiring process went so much faster when we just simplified it and said, show us you doing this task. The best way to tell if somebody can do a task is not to talk to them. It's to look at their work. I think that's, uh, I think something all CEOs know inherently and all high agency people know, but a lot of young people in the workforce are trying to figure out, all right, let's go to our next topic. Training data. Ali, you do this for a living at MicroOne, and Mikey, you're very focused on the music space, so I got a lot of questions about this. How much training data is left, Ali, in the world that hasn't been scraped, open, crawled, etc., or licensed now? Because we're entering this age of licensing where Reddit, Quora are building really nice businesses. Obviously Elon bought Twitter in order to have proprietary training data, and that's why Grok is so good at real-time queries. Is there any data left? And how are these large language models or verticalized models creating data? And it's a bit of a softball question since this is what MicroOne does for a living.

Yeah. So, so there is, there is no data left. Um, you know, there's a, the internet has obviously been, been trained on, and I think there's a large-scale kind of equivalent datasets that, that have similar quality. So training on them actually doesn't really meaningfully improve models. So, you know, the way we look at it is we're in the phase of post-training where net new data is required, and that net new data is created by humans, oftentimes human experts and subject matter experts in specific domains that are creating very complex, high-fidelity datasets to continue improving model capabilities. And I think the route that, you know, foundational models companies are taking is they're trying to improve in the highest economically valuable domains. And they try to improve in, in with data structures that are emergent, which means, uh, you know, improving not only in that domain, but also improving the model generally.

If you're not listening to your customers and iterating based on their feedback, your startup will fail. So if you're drowning in feedback from support, reviews, calls, it's time for Interpret. Interpret is the customer intelligence platform that helps you turn random comments and feedback into actionable insights that help grow your company. Their AI system is already trained on your business and your product. It then reads all your support tickets. It reads all your reviews, call transcripts, and surveys. Then their agents actually update your team in Jira, Linear, Zendesk, and Slack. So this feedback doesn't just sit in some email or messaging thread, but it actually helps keep your team moving the ball forward. Find out why Canva, Notion, and Perplexity are already using Interpret to stay on top of what they need to fix or build next. So if you're ready to turn feedback chaos into customer intelligence, head to interpret.com/twist to book a demo and see it in action. That's interpret.com/twist. Give some examples of the top, you know, high-value verticals. Obviously people have watched, we have an investment in in a company called TaxGPT that's doing tax law specifically, which is granular, which is detailed, which you need expertise in, which is always changing. Obviously tax laws change every day globally in, you know, hundreds of markets and thousands of different little niches. So I'm assuming accounting and legal are obvious ones in there, but, and maybe coding. What else is in this list? What are your clients asking for, generally speaking?

Yeah, the three main categories are finance, medical, and legal, and coding, You know, is a, is a very close fourth. Those are the ones that we sort of focus on based on the demand that we, that we see. I think that the, the current state of the market is very, very much still in this phase of getting the models to be really good at question and answer pairs, like really complex questions and very complex domains. But we're now starting to enter the phase of models actually doing things in those domains. And I think tax is actually a really good example. If you, if you think about like a, You know, what is a question-answer pair in tax? It may be, you know, giving your income of the year and then asking for tax advice on like, you know, purchases you can make before end of year. And the models do, I think, a pretty good job at answering these singular questions. But where models fail at very much so currently is when they have to answer sort of like multiple questions in a row, which is equivalent to saying they have to sort of do things. And the simplest way to think about this is if you assume that there's any given domain that has, let's say, 90% accuracy in answering any given question, any arbitrary question. If you, if you take that and you have to answer multiple in a row, let's say 5 questions in a row to actually do some actions and you do 0.9 to the power of 5, you're going to very quickly kind of go to zero, right? The accuracy is going to compound, uh, converge towards zero very quickly, which is why training data for tasks that have these sequential actions is much harder. So one of the environments that actually we're creating is a, is a tax environment specifically for California W-2 taxes. And as you can imagine, there's lots of, you know, steps to filing someone's taxes, such as collecting input data from them first, making sure you actually have the right input data. There's some partial rewards for that. And then there's a conversation that happens around tax advice and a bunch of steps in between. And then ultimately the tax is filed and each of those steps, you know, they depend on each of the previous states and the previous states need to be accurate for the, you know, for the following states to be accurate as well.

What are you call this career where a tax expert who's 10 years into their career can say, you know what, I don't like clients anymore or customers. I just want to answer questions and, you know, be heads down talking to an LLM and making it better. What is the career path that this is now, Ali?

I don't think it's truly exclusive to like not liking clients and wanting to just train models. I think the way we see it is these experts do both. The category is called AI training, you know, AI training experts or human data experts. In fact, one of our competitors actually just came out with this study that there's over 200,000 people doing these type of jobs now, I believe just in the US, might be internationally, but I think it was just in the US. And so there's a lot of folks that are, you know, spending 15, 20 hours a week of their time training models. And I think that the way to frame it is they are getting paid to do things they're already really good at, to— and they're getting paid often more than their day job in almost every case, actually. To train models that will help them do their job in the future. And I think that's a nice way to look at it.

HDEs, human data experts. I love it. And they could be doing this freelance, you know, as a side hustle, be doing it very quietly. They maybe they work for a Big Four accounting firm or whatever. And then on the weekends they get, you know, $200 an hour, $300 an hour work to do this kind of stuff.

Yeah, exactly.

And I don't think most people are aware of this fact that that this is what training data has moved to. Wade, you have, I'm assuming, hundreds of thousands of people every month architecting workflows which you can then have as proprietary training for the Zapier god operational— I don't know what you call this person, but there's some air traffic controller in the Zapier brain. Tell us about about that? Are you building the Zapier brain of operations based on all this, you know, ticky-tacky stuff that my team's doing to take data from one repository, get it into Slack, and then bring it back over to Notion and, and move it out of email into structured data? Tell us about your thoughts on training data.

Yeah, I think, you know, any company that has, you know, uh, proprietary data, um, they're trying to figure out, hey, can we, can we use this to provide like a better product experience, a better application than the foundation models can directly provide themselves. And so, you know, in the case of us, we're sitting on 14 years of workflow data and as a result, like we know how to make these things reliable and when things break, we know how to fix them. And, you know, we can use that data because very few other people have access to the corpus of information we have. And so we can use that to to improve the experience. And, and that's kind of, you know, I think what most companies are doing that have these like large corpus of bodies are trying to figure out how do we build a better product experience because we have access to this and no one else does. To Ali's point, like all the public information has been, been, uh, soaked up at this point in time, but, uh, the private stuff is where a lot of the alpha is today.

Have you been approached by the Sam Altman's of the world, the Claude's of the world, to take your dataset and they license it from you for for $50 million a year? And how do you think about— because I have had startups who've had this kind of internal discussion in board meetings. Hey, should we keep this proprietary? Should we license it? Obviously Quora and Stack Overflow, they've gone the licensing route. Do you think, hey, maybe I should just license this to an LLM?

Well, you know, I think the— it's, it's no secret that I think the foundational models are, um, have an insatiable appetite for, for training data. And so they're going out and spending large amounts of money to get access to these proprietary tools. I think our approach is, you know, we'd rather keep that for ourselves. But I do think there's benefit to releasing benchmarks and evals and other tools that can help you get a sense for how well these models perform against these tools. And if you make those public, you'll likely find that the foundation models will start training on those things. And, and that can be beneficial to you because you're getting some cooperation with everybody to improve on a certain thing that you actually are like, I, I wish the models were better at this because if they were, I could create a better experience for my customers. Uh, and so I do think a lot of companies are looking at how they release benchmarks.

And are you, uh, integrating yourself into like OpenAI's— I don't know what they call them now. I think they were originally called apps or integrations.

Connectors.

Yeah, connectors. How do you think about OpenAI's connectors as a competitor or a collaborator for Zapier?

I think it's a pretty important collaborator. So, uh, we're a connector inside of ChatGPT. You can use us, uh, as an MCP provider inside their agent builder.

Explain what that is, MCP, for people who don't know. Yeah.

So MCP is Model Context Protocol. You can almost think of it like APIs for, uh, agents to talk to each other. It's probably like a, a simple way if you're unfamiliar. It's really important that these models have access to tools and data. Uh, if they don't have access to them, you're not going to get a great answer. So for example, imagine you're running a, you know, an insurance shop like in the middle of the country and you're like, you want to ask ChatGPT, can you generate a marketing plan for me? Well, you'll probably get a marketing plan back that you'll go, well, it's, it's decent, but it knows nothing about my business. And so I'm not actually going to be willing to use it. However, if you say, hey, you know, ChatGPT, let me give you access to the last 10 campaigns I run. Let me give you access to the types of customers that I like and the types of customers that I dislike. Let me give you access to my CRM data that has all that closed one data, that closed loss data. Let me give you access to this, that, and the other. Now, how about you go try again and generate that marketing campaign for me? And you'll find that you get just a much better output after giving access to your data, your tools, and things like that. Uh, and, uh, so, so Zapier plus, you know, uh, you know, a ChatGPT or a Claude or a Gemini is a pretty powerful combination.

Yeah. And that means they have to have an account and if they get value from it, they put their credit card in and you get a new customer. So it's a win-win. Uh, Mikey, you have the most, uh, difficult, I think challenging task, which is the music industry. They're pretty, um, uh, protective of their IP. And, uh, very litigious, obviously. And then it's, it's super valuable. They also have to contend with their artists. And does Mark Knopfler want to be part of this and have people— you know, the first thing I do whenever I use one of your tools or one of them is say, hey, make me a Mark Knopfler song, because Dire Straits hasn't been making albums since On Every Street in the early '90s, or yeah, in the '90s. So I'm desperate for a Dire Straits album. Um, how do you think about training data? Are you hiring musicians and experts to make riffs and play the blues and then training your data with proprietary data like that? Or is it really you just have to make relationships with all the labels and, and the artists and that's the mode in your business? Because there are these sampling websites, right? We've— there's a number of sampling websites that let you buy samples and build songs. So Explain to us the game on the field right now. Splice, right? That's the name of it, Splice, because that's all open, that's all royalty-free, I believe.

I don't want to say one way or the other because I don't know exactly how Splice does it. But, um, yeah, I mean, actually the funny thing is everybody here is actually a musician, or almost everybody here is actually a musician. That's what I was talking about earlier, that like the, the mission alignment is ultimately what gets people to want to come work here instead of, um, you know, getting a lot more at OpenAI. I think there's— look, there's a lot to say. I have always said, though, that yes, it is an organized and sometimes litigious industry. It's not for people without a thick stomach lining. But at the same time, there's a lot more to do together than fighting one another. I've thought that since the day we started. This. And I understand that like music and technology don't have the most straightforward past. And so there's a lot of, you need to show and not tell. But like when you do actually, people get it. And they're like, AI can be a tool to make the music industry bigger, to deepen the artist-fan engagement, to make amazing tools to let people make music that they have never been able to make before. And once people see that, they actually kind of get on board. This was recently put to me by someone as, um, we're the GLP-1s of the music industry. Everybody's on them, but nobody wants to talk about it. And, uh, you know, I think slowly, eventually people will start to kind of, um, admit that they're, um, using and enjoying the stuff that we make.

I can tell you I am on GLP-1s. I have for 4 years since my friend Kevin Rose and Tim Ferriss talked about the off-label use of them, and I lost 43 pounds. And you can go to ro.co, which is where I get my GLPs, and I'm now a, uh, a spot— did you guys know I'm sponsored by ro.co?

We do now.

I'm an official spokesperson. Me, Serena Williams, and Charles Barkley, 3 exceptional individuals at their craft. So I literally— it's— I've done 2 of these. It's not an influencer deal. It's like being a brand. It's like being a celebrity. So those are my— that's my new celebrity endorsing. I'm an endorser. I mean, it's my endorsement deal. But they are using it and they're making great music. And Mikey, they don't need to be— people don't need to be precious about this tool, do they? Because a lot of the great music of the century has been based on sampling some incredible David Bowie song, some incredible hook. I mean, Kanye West, Diddy— no, Diddy, you know, whatever. I was in jail and this Diddy documentary is pretty brutal. But, you know, he was famous for sampling the, um, Sting song "Every Breath You Take" in his tribute to Biggie Smalls. This is not new to the industry, and they actually are particularly good at figuring out how to create compensation already for sampled music.

Yeah, I think, I think that you're right, this is not new. Um, it was with sampling actually. If you Go back to the— another huge technological change in how music was made was just digital production, that you used to need to go to a studio and make music on a really big expensive console, and now, you know, you can have a 13-year-old bedroom producer doing it on their laptop. And at the beginning, people said that's not real music too, and that led to new genres of music, that led to an explosion of music, that led to, I think, a very positive force. Um, I think the thing that's different is just, um, things are happening very quickly now. Like, everything seems to happen— you know, you said it feels like 10 years ago, um, when OpenAI was doing everything they were— or when Meta was doing everything they were doing during the summer, and that was 6 months ago. And so, um, uh, it is a very dynamic time, um, but the best future is built, um, in partnership and not fighting one another.

Yeah, I remember when I was Coming up in New York in the late '90s, I used to play basketball with a DJ. His name was DJ Clue, and he made mixtapes, and he started by selling them literally at train stations or from the back of his car kind of situation. And basically sampling and remixing started as like just selling tapes, you know, in Harlem, on Canal Street, etc., and then it became a big part of the industry, uh, and you had breakout artists who who were part of that. Has there been a breakout artist from Suno yet where they use the tool and they broke into the music industry building AI-first music?

We have. It's still very early, of course, but, um, as an example, um, there's an artist, uh, with the name Zanaya Monet. Um, I don't know if you saw that, um, but it's a, it's a lady named Nikki who, um, is an incredible poet who writes from like the this really deep emotional place. She puts her stuff, she puts her lyrics into Sunoo. She makes these songs. Those songs resonate very deeply with people. And she got signed to a record deal, like a real record deal, not just a distribution deal. And I think that's amazing. Like her music is clearly resonating with people. And to me, that is not depersonalizing it. There's a real person behind it. She's incredibly talented. Here she is. So I think like—

She got the follow.

And this is not her. These are all AI images. But if you go and you listen to the music, it's like, it's really touching.

Amazing. So this has already started to happen. And this is when you know something's happening is when I was talking to Tony Hinchcliffe from Kill Tony, the people at the major artist representation firm firms, like agents, are now using Kill Tony as where they source the next wave of talent. Just people who've gone on and done a minute of Kill Tony, somebody will get to them before the Kill Tony episode has even come out. And so this is what's going to happen with Sunoh. Yeah, you're going to eventually be the place where people are looking. And is there like a public repository the same way, you know, Sora has like, here's a gallery of, you know, things that are trending? Is there like a trending on Suno somewhere where people can look at the trending tracks? And do you host the stuff, or do people just go on DistroKid and my friend Phil Kaplan's company and then put it on Spotify? Tell me about the bridge. And is that part of your plan, is to build that bridge into the, you know, mainstream industry?

No plans right now. Uh, we do have a trending page on our, on our website. Humbly, it's a little, it's a little crude still. Um, you know, it's, it's not been the the biggest focus. But no, I mean, stuff is there and gets a lot of plays, and people love when their stuff gets a lot of plays. I think it is like a basic human desire to want recognition and admiration for the fruits of your labor. And so I don't think that's something to be ashamed of, that like, hey, I want to actually like get to the Tsuno trending page. Another artist, Oliver McCann, with the name I'm Oliver, has, I think, the most played Zuno song ever. And because of that, also got signed to a record deal. And so, like, it's stuff like this that is actually amazing. You would never have found him or Zanaya Monet without giving more people the ability to make music.

Amazing. All right, let's go to a bigger, broader issue, which is acceleration versus deceleration. Bernie Sanders seems to have joined the AI discussion. And listen, it's— this isn't going to be a politics show, but AI is having a major impact in the world. We've never seen a technology this powerful, and we've never seen a technology advancing this quickly. Everything we saw last year seems, you know, incredibly basic and crude compared to, you know, the products we're seeing this year. And the same thing will happen in the next 6 months. It's only accelerating. Here's a— and I'm assuming you all saw— did everybody see the clips of Sanders? I'll play a couple of them here, but I think you guys have been seeing this because it's trending. All right. He makes basically 3 arguments. The first is billionaires, which in the socialist mind are terrible things as opposed to a capitalist or a democracy. Everybody should be able to have their own path to success. Here's the first clip. He's blaming all this on billionaires who want more wealth and power. Play the clip.

Surprise, surprise, it happens to be the very wealthiest people on Earth. People like Elon Musk, Jeff Bezos, Mark Zuckerberg, Peter Thiel, Bill Gates, and other multibillionaires. So here is a very simple question I'd like you to think about. Do you believe that these guys, these multi-billionaires, are staying up nights worrying about what AI and robotics will do to the working families of our country and the world?

So that's the framing, that's the setup. I know all these people, I can tell you they actually are worried about this and they're thinking about it. But that's the sort of the setup, and not wrong that every important technology company and person is focused on this. Um, but here is, I think, where he makes a pretty salient argument, which is that robotics is going to create job loss, uh, specifically the robotic space. We'll play that clip and then we'll talk to the panel about it.

What will AI and robotics mean economically for the working class of this country? Well, Don't listen to me. Listen to the people who are developing those technologies. Elon Musk recently said, quote, "AI and robots will replace all jobs. Working will be optional," end quote. Dario Amodei, the CEO of Anthropic, warned that AI could lead to the loss of half of all entry-level white-collar of jobs. Question: if AI and robotics eliminate millions of jobs and create massive unemployment, how will people survive if they have no income?

So there's the setup, Ali. Uh, what does he get right? What does he get wrong? And how do you think our industry should reply to this I would say is a somewhat valid, uh, uh, concern.

Yeah. So I think what he gets right is yes, AI and robotics will likely eventually replace all jobs. But I think the, the, you know, the way to think about that is it replaces the current state of jobs and makes jobs a lot more fun for humans. I think we're already seeing that, right? There's certain jobs where You, you no longer have to do the very tedious parts of the job and you can do the more strategic kind of creative parts of it. So I do think he gets that part right, but I actually think that's a, a good thing for society. And, and, and eventually if we get to the point of, you know, jobs are optional, that, that also is, is a good thing because it means that there is, uh, some notion of universal basic income and you actually don't need to work and you can choose to work if your job is you know, fun in your, you know, in your view. I think what he gets wrong is this, you know, this, this, this constant argument of billionaire is bad just innately. You know, the framing that I would take is the reason why certain folks are billionaires, because society has kind of decided to allocate those funds to them because they are good allocators of resources and, and those funds. And I think you can kind of make the analogy to if they are innately good at allocating resources, they will likely be good at allocating, you know, compute and, and data to train these models in a, in an efficient and in a, in a human-aligned manner. So that, that, that's, that's sort of what I would say. And I think the, and I, I agree with you, Jason, that, that I think these folks are thinking about the working family, are thinking about the consumer. Because they are— I mean, they're literally building products for those types of folks. I mean, Elon is a good example of almost, you know, all his revenue from all his companies is selling directly to the working class. And he needs to serve the working class by the goods and services that he builds for them.

Yeah. Wade, when you hear this rhetoric, being our salty dog here with, you know, now in your second or third paradigm shift in the tech industry, What's your take on, you know, the timing of job displacement? I'm not going to call it job loss. I think it's jobs are being displaced. They always have been. But we can't argue it's not going to happen at an accelerating rate because as we started these discussions, it is happening. People are becoming much better at their jobs. But I think we're all hiring here. So what's the ground truth here? Because, you it does seem like displacement's gonna happen in certain jobs, the tedious ones, but it's also happening in white collar jobs. You're seeing more efficiency in operations. And we talked about how Zuckerberg and Google don't believe in product managers. And you yourself said, hey, we need people who embrace these tools while still having some experience. So unpack his position. What he's got right? What has he got wrong?

What we're seeing is just the continuation of inventors building things that make society better. You know, if we rewind the clock hundreds of years, most of us are agrarian farmers. Uh, and I think most of us would agree that life is better now than it was, you know, 500 years ago. And I think in the moment it's very difficult to understand what the new opportunities will be. It's much easier to look at what we will lose. But if you take this discussion here, All of us have talked about the new jobs that are being created. All he's talking about these new like data training jobs that are coming up. Mikey talked about these new people who are now able to create music. You know, we have folks that were living in cars who now have automation agencies. So like with new technology comes new opportunities, new jobs. And I get that it's understandable to be like fearful of what might lose. But I think we have to understand that for For centuries now, humans have come up with new ways to do things that are economically useful, whether they're providing goods and service, whether it's entertainment. And so I think it's a little bit short-sighted to say that this stuff is bad. I mean, we've got— ChatGPT alone has 800 million weekly active users. Clearly, society loves these products that are being made, and no one's forcing anyone to buy them, but yet we choose to use them and buy them.

Mike, you're— literally, we just in the last segment talked about all these new careers that are going to happen because instead of, I don't know, 1 or 2% of society being able to make a production-ready track that somebody might really enjoy listening to, to— I think your tool probably makes it possible for, I don't know, everybody to make something production-ready, or I don't know, a third. What, what are your— what's your take on this decel approach that Bernie Sanders is advocating for?

I understand where he's coming from. I think it's a little misguided. I'm not sure I see the connection between what he said and stopping data center production. And maybe there's even a case of like, if you stop it, who do you think is going to buy up all the current data center capacity? And I think You know, so to me, these things are not connected. Yes, I will also just say the same thing that everyone else said. The fact that an ad hominem attack on people like we don't like the messenger, so the message must be bad just seems deeply wrong. But yeah, I would just think about like the world that we are trying to build probably needs a lot more power and probably needs a lot more data centers. And if we don't do that, Um, everybody is worse off. And I understand his concerns around, um, maybe things need to be a little bit more distributive, but if we don't build it, there won't be anything to redistribute.

Yeah, here's a good clip of his solution to all this. Um, and there's another interesting side quest we'll go on about, you know, what this is doing to kids and people being less social, but Let's hear Bernie's solution, which is to get rid of the data centers and the energy.

The unregulated sprint to develop and deploy AI. This moratorium will give democracy a chance to catch up with the transformative changes that we are witnessing and make sure that the benefits of these technologies work for all of us.

I mean, this is, I think, where he goes off the rails. We're going to stop the data centers, Ali, and then if there's no data centers, then there will be more jobs and people will not lose their jobs. It's the craziest possible solution. There are other practical solutions you could think of if in fact driving jobs are going away and you're concerned that's going to happen at a really violent rate. Rate and we're gonna have 5 million truck drivers and cab drivers and Uber drivers and DoorDashers suddenly out of work, you could create a tax or a license for those. Uh, not necessarily saying I'm advocating for them, but there would be other solutions. So what are your thoughts on if, you know, there was a license for a robo-taxi that charged an extra tax and then that tax went to retraining funds, etc.? Any thoughts there, Ali?

I think that that's sort of a similar incentive to actually reduce innovation. I think if we're taxing robo-taxis, we're kind of not incentivizing them to spread like they should, in my view. And I think if we, you know, if we follow what Bernie is suggesting as a solution of basically a full stop on AI so that he can kind of catch up to understanding it and the government can broadly catch up to understanding the systems, we will simply lose the AI race. I think if we fully stop and really implement the solution, China will win the AI race and we will be forced to use Chinese models, which I think we certainly don't want US data and US infrastructure to depend on AI models in China. And I think the important thing to do is obviously, you know, think about the future and try to mitigate the risks of of job loss. But as importantly, you should kind of analyze the current state, which is there's, you know, as we said earlier, there's a lot of new jobs that are being created and a lot of existing jobs are simply becoming more fun and creative. And I think, you know, if I were to ask Bernie, like if he knows about the 200,000 AI training jobs that have been created in the US, he probably doesn't know about that fact. And I think we need to kind of educate folks that have this point of view on the current state of the world, which is very beneficial to humanity.

Yeah, Wade Foster, it would be absolutely extraordinary if 30,000 people didn't die in car accidents unnecessarily a year, and if the price of Ubers and DoorDashes went down 50% or eventually 90%, well then people, uh, are going to get that benefit as well. Your thoughts as we sort of wrap up the segment?

Yeah, I kind of just come back to these products are very popular. 800 million people use ChatGPT, and that's just that product alone. Tons of folks use the products that we're, you know, each of us are founders of. You talked about Uber and, um, robotaxis and Waymo and like the, the, um, life-saving that they— that can cause. We're not talking about like, you know, uh, new drugs that can solve, uh, you know, diseases and all these other things that I think are very positive. And so I do think it just feels very short-sighted to say, hey, we're going to pause data center construction because of that. Um, I'm not sure what the ultimate solution will be. Like, I understand that, you know, if you're, you're in a policy maker, you got to think through like the societal ramifications of the— and there needs to be a debate around that. But this, this solution feels, yeah, pretty short-sighted, honestly.

Yeah, it's a dumb solution. I, I've been giving this a lot of thought, obviously, and we've talked about it on All In for the past 2 years. I've taken the position that the job displacement's happening faster than we think, but I don't think de-selling is the solution. I think actually the rallying cry for our industry should be able to communicate 3 areas of focus that would benefit Americans, specifically young Americans. The 3 most overpriced, uh, items right now in America are education, healthcare, and housing.

All highly regulated.

All highly regulated. Uh, and all of them could be disrupted in a major way by AI. Robots building homes, healthcare becoming super cheap. Um, Daniel Ek is working on reducing the cost of these 3D body scans to just a couple of hundred bucks. What— and obviously education is going to be dramatically impacted by this. Why would you spend $100,000 on a degree when you could teach yourself the same tools, uh, and become amazing at them without spending $100,000 in higher education? Why don't we just, as an industry, show the world specifically Americans. Hey, we can really lower the prices of those three really acute things and make it easier to own a home, easier to educate your kids and get ahead of this. And then, uh, Mikey, you know, live longer and lower the cost of healthcare. All of these things are going to happen. I think we have a communication issue in the technology industry with the Democratic socialists, and we just need to bridge that gap. Hey, how would it sound if healthcare was free or close to free? How would it sound if homes got built faster? And how would it sound if, you know, education was free for everybody and you could get a really great trade skill, uh, very quickly using these tools?

Can't follow that. Hard to argue with any of that. No, I, I think, I think, um, tech has a pretty bad, um, communication problem in general. I, I know our corner of the universe of it, but Uh, I, I think that the, the best, like, evidence of that communication problem is the fact that he hates tech billionaires. He didn't say anyone outside of tech who's a billionaire, and there are plenty. And so, um, um, yeah, I don't know, I have nothing else pithy to say. You, you kind of nailed it.

Well, here's one where I think is in your wheelhouse, because you think about people having joy from making and experiencing music. And we do have, uh, maybe a generation, the COVID generation, or the, you know, permanently online generation. He did throw in a weird curveball into this, uh, decel argument and banning the data centers that I thought was interesting and worth unpacking. Let's play this weird curveball from Bernie.

3, something I worry about a whole lot, and that is that millions of kids in this country are becoming more and more isolated from real human relationships and are getting their emotional support from AI. Think for a moment about a future when human beings are not interacting with each other and are spending virtually all of their time with devices instead of people. Is Is that the kind of future you want? Well, not me.

Michael, let's unpack that one, because this is definitely true. People are, you know, and they said this about every generation. For us, it was we were spending too much time watching MTV. Then we were online too much. Now it's TikTok and Roblox and character AI. It is true that people are getting isolated and that people aren't leaving their homes. And I think it's a double whammy of the addictive nature of this technology. And COVID just making people into homebodies. What are your thoughts, Mikey?

A lot of thoughts. I'm, I'm a father of two small kids, and so, um, I look— though my impression is that most of this isolation does not come from AI, it predates it. It's a lot of social media. You can say there are pros and cons of social media. Um, I think I do somewhat, uh, reject the idea that AI can't be actually a really force— good force here. In the same way, imagine you can give the entire world a median competent therapist, um, with AI. Like, that's probably good for some of the mental health crises that we face in this country. Um, and there's lots of stuff to do about it. To me, it's just like somewhat incongruent to say that the correct solution to that problem is to ban AI. You're probably going to make things worse. Um, I'm all for though, like, I think there's a really important nugget of truth in what he's saying, which is that human fulfillment is really important. Probably working is an important part of being fulfilled. You know, something we say at Suno is we say fun is underrated. Um, and, um, there's lots of people working on productivity. Uh, the, the other guests here, uh, are two important ingredients there, and I think there's actually not enough time in this ecosystem that we all play in spent trying to promote human flourishing, trying to make products that give people fulfillment. We like to think we are a very small part in making people smile. There's no illusions. I know that Suno is not curing cancer, but it's actually kind of amazing when you see some of the stories and you know how many people you make mile every day.

And you have young kids too, Wade? You have under 5-year-olds, under 6-year-olds?

Yeah, I have, uh, two, two young kids.

How are you thinking about their technology use and then maybe interacting with super sophisticated AI instead of talking to their friends? Uh, and then obviously social media. I mean, I have two 9-year-olds and a 16-year-old now, so I'm part of the social media, trying to keep that floodgate. My daughter desperately wants to get on social media and we've been able to push that off. But at 16, I think maybe okay to start dabbling, or maybe 17, in the Instagram world because her friends are all on it now. But we gave her her phone, I think, at 14, and iPads, you know, since 5 or 6 years old. But you got to really limit this stuff. How are you thinking about raising kids into this crazy addictive soup of character AI, etc.?

Yeah, well, I'm fortunate my kids are young enough that, you know, this is not yet a problem. But I am starting to think about, like, you know, the kind of, you know, sort of environment we want them to be raised in. And I do think, you know, it's important that, you know, they have— that they are involved in the community. And so we want them to, you know, be in person at school. We want them to be in dance and gymnastics and sports and music and all this other stuff because they think they get to interact with other people and they get to learn those skills. And I think it really is like an important skill to just like to be able to interact with other people and interact with society and sort of run into conflict and run into disagreements and know how to navigate those things in ways that are respectful and polite. And so I do think that that is a really important thing to do. And so I suspect that as we, you know, the kids get into that age group, we will probably find ways to keep encouraging more of that and limit, you know, time with the phone where you're just like on social media and some of these other tools. It doesn't mean that we'll like get rid of those things. I just think that that's something we'll have to, you know, teach them how to use responsibly at the end of the day. Yeah, I think that's kind of The challenge for every generation of parents is that there's always something new and you got to figure out how to, how to introduce them to the good parts of those tools and, um, you know, teach them how to navigate the things that are, that are more challenging.

And you know what, you have to teach adults how to do this too. When we were in Vegas, we all went to dinner, uh, me, Chamath, uh, Freeberg, and a couple of other folks. And, uh, I said, hey, everybody stack your phones. And we all put our phones in a stack on the side of the booth. And I said, first person to touch it pays for dinner. Let's all just talk and be present, you know. And same thing with poker, you know, we're playing poker sometimes. I was at a poker game last night or two nights ago, and I instituted the phone penalty, which is every time you're on your phone and the action's on you, you put $25 into the pot and then it doubles each time. And it got to like $200. People started to put their phone down and experience experience being with each other. You have to actually create these moments. You have to be thoughtful about it. And it's the job of parents. But look at this list of, these are 50 Gen AI properties by monthly visits. And I think A16Z shared this recently on social. Of these 50, it looks like 3, 4, 5, 6, 7, 8, 9 of them are character AI, you know, Spicy Chat, Candy AI, Juicy, You know, all of these character-based ones, these are particularly addicting, especially if you are socially isolated already. These are the ones I don't think kids— these things should be 16 and above or something. You do not want your kids getting into a crazy relationship with an AI character, uh, long term. You got to create these moments. I started reading at night because I'm so addicted to You know, TikTok and Instagram and mostly X and just, you know, trying to read every night for the last hour or half hour is such a better win in terms of getting yourself to sleep. Uh, and so this is going to take thoughtfulness on parents. All right, listen, we've been going for 75 minutes. I guess we could either save the year of the agent or we can do it lightning round. You guys got 10 minutes to lightning round it? Okay, we're gonna lightning round it. Wade, agents were supposed to be here this year and we were supposed to all be hiring an agent to do HR, to do customer support, and not have to do these functions at all. It's still kind of Copilot-y. It's still kind of vibe code-y. When do the agents actually start working and we can just give them a job and not have to be copiloting it? When can we just give them the controls? And what's the holdup? What are the blockers?

Well, I do think, you know, the year of the agent is probably more like the decade of the agent. But I do think this year was, was chapter 1. We saw some pretty, like, positive advances. I think the real challenge that we have with an agent is that the way in which folks want to use them is they want to use them for these very complicated tasks. And any time you want to use an agent in a complex production case, use case in any sizable industry, you're asking that agent to, um, chain along multiple decisions. And if these agents have even, even 90% reliability on any given task, well, if you're chaining 10 tasks in a row, the end result is that by the time you get through all of those tasks, the end reliability is quite low. And so I think that's what most folks struggled with this year, is, um, that just sort of reliability hill that they have to climb. And so what we've seen a lot in practice is in our customer base is combined, and we have agents and workflows. And so oftentimes folks are combining these two concepts together. You know, a workflow— the way I distinguish the two, by the way, is that a workflow is very deterministic. It's set of nodes. It's if this, then do this, then do this, then do this in a very structured way. Whereas an agent has a list of instructions and tools and data associated with it, but the agent gets the decide how it goes and performs those tasks. And if you put an agent inside of a workflow, what you can then do is turn that complex down, that complex task into a set of smaller subtasks. And in doing so, many of those can be just deterministic sets. And then you insert the agent only in the place where you truly need an LLM to make a decision, to categorize a thing. Um, and as a result, you end up getting much more reliability. And so that is what is really working in production today is this like agentic workflow, this sort of like middle ground that is being quite successful. Ultimately, I think what is going to enable us to deploy more of these agents in a fully autonomous way is one, you're going to have better models, stronger reasoning, better data. You just need more access to more data and the tools, access to tools. And so there's a whole whole set of environmental things that are required to get us to eke out those reliability gains over time. Probably the last and maybe most important thing is there's just not a lot of examples of these. They're quite difficult to build. You still need humans who can think through these concepts, break them down into systems, and just not many people have this skill yet. But I see it accelerating here at the end of the year. And so I am optimistic for, for how much we're going to see more fully autonomous agents, um, next year.

Wait, Wade Foster, is there a domain that you think is most likely to have like the first killer app of agents?

Well, I mean, the ones that are talked about right now are coding agents. The coding agents are, are incredible. Uh, you know, I think customer support is also having quite a bit of success. But, um, we have seen these agents deployed in every domain. Uh, in fact, I think outside of, uh, developers and customer service, most functions are not being as creative as they could. I think there's tons of fantastic use cases for marketing, for sales, for finance, for HR, and every single one of these functions we have internally agents running that are operating tasks for, you know, critical, critical workflows. But when we go spend time with customers, many of them are still not sure what's even possible.

Which one impresses you the most currently, Wade, that your team built? You have one that you're like, oh my God, as a founder of this company, I'm just over the moon about this one because it's so annoying and it's just to have to do it manually, or it's just so great to do it with an agent.

One of my favorite use cases right now are call transcripts. So you might use something like Gong or Granola or many of these like tools that will help record a transcript of a call. And there's so much trapped value that happen inside those conversations. These are— they, they happen many times a day inside your company, and usually they are very ephemeral. You get off the call and all that context is lost. But if you're doing a good job, you can capture that context and then do very interesting things with it. So if you take, for example, a customer, a conversation with a prospect or a customer, you can do a handful of things. One, you can turn that into a coaching agent that coaches your sales rep on how to do a better job on that call. 2, you can take questions that the customer asked and you can generate FAQs based off of that. And so now you can publish those FAQs on your site. If you're talking to an existing customer and they're talking about why they like your product, you can auto-turn those into case studies.

Ooh, testimonials.

Love that.

Testimonials are great. If they talk about the value of your product or the use case in a novel way that you haven't heard before, you can generate new landing pages. That target specific keywords or topics that maybe you don't already have on your site, then you can generate ads against those keywords. And so now you could be generating those in a semi-automated or fully automated way. And we could go on and on. But the result is that this context of a conversation that you are having has so much interesting nuggets that when you feed those conversations into an LLM and target those to generate specific artifacts that you would want to use in other ways, you can get a huge amount of value and you can cut out so much in-between work. And the reality is most companies, they just don't do it because it's like, it's just too, it's just too burdensome to do these types of things. But now it's very simple to set these types of agents and workflows up.

Yeah. And you just have to slow down to speed up. We have an agent that takes emails and incoming applications from founders. We do two things. We have our Athena assistant, go to AthenaWow.com and invest in the company, uh, and you can get a couple weeks for free. We have them taking the inbound applications for funding at our fund and at Founder University and just making sure we fill everything in. And then also we'll take a deck and normalize it and create summaries. And when we do our first meeting, like I have the first meeting, Ali, you did with our team because we started this database 3 years ago, and I'll pull the video at some point and show it to you so you can see your pitching skill from 3 years ago, but we can basically learn from those videos. So now in year 4 of recording our first calls with founders, putting them into the database, I did a thing that was really interesting. I, I had a company I couldn't remember that had pitched me on something, and I said, tell me, because Notion built-in AI, and then I connected the Notion AI to Slack, and Slack's internal AI isn't very good compared to Notion's. And when I said, hey, tell me about this company that was doing this Agentic thing, speaking of agents, that we invested in, and then were there any others that we didn't invest in? It gave me a summary of all the companies that we passed on investing in. And so I was like, wait a second, can we have an agent going out and finding out who we said no to, we passed on investing in, who went on to go to Y Combinator or Speedrun from A16z or Techstars? Stars or raised a Series A. And then we could actually automate the anti-portfolio and then say, why did we screw up? And so we have all that proprietary training data. And that's, you know, my big mission in 2026 is to make those agents actually do the full cycle of a venture capitalist job and see how far we can take it. Ali, you have some thoughts on this as well, because you're obviously building training data, but you need tools. Automate the building of training data. So take us through your thoughts on it.

I think Wade Foster said it really nicely, which is we were, we are in the chapter 1 of agent buildouts for enterprise and otherwise. And I look at the bottleneck, uh, you know, putting aside the model capabilities improvements and the foundational models improving overall, I think the two biggest bottlenecks are this notion of like MCP slash API connections for softwares and workflows to, you know, connect together. And I think that the second one is data and specifically specifically contextual eval data for enterprises. I think, uh, enterprises have not yet really thought about or spend really any efforts on describing the functions that the agent needs to have, quantifying what it means for the agent to do well in those functions, and then actually evaluating those. And that requires a, you know, ton of human data that is very niche to any given enterprise workflow. And I think 2026, what I'm sort of predicting is that very large portions of enterprise budgets will end up going towards human data and contextual evaluations. And it's almost, I think eventually it'll become sort of equivalent to like an engineering function at an enterprise where like there's an evals team that spends a lot on gathering domain expertise as well as niche enterprise kind of workflow evaluations. And I think that the The reason for this is, um, if you're building some sort of probabilistic software, you cannot check whether the thing works or not. It's not like a binary kind of like typical QA engineering type, um, workflow. It's more so you have to check, does it work well? And what are the parameters under which I'm defining what well means?

All right. Listen, and this has been an amazing first episode. Here's your time capsule. Some founders are so mercurial and and effervescent that they'll pitch us from completely inappropriate locations. Can you imagine this, Mikey? You're a founder and you decide to pitch a legendary angel investor and his firm. Well, here you go. Play the tape. Look at this guy.

In front of an ice cream shop.

Yeah. Shout out to Laura Lake, who took this call, who used to work for me. Look at this. Microwave. He's had a Jenny's ice cream.

What were you thinking, Ollie? Oh, I can't believe you actually pulled this off.

Okay, there it is, folks.

All right, you can—

we're not gonna put any sound, but I'm gonna watch this after the show.

That's—

what date is that from, Producer Oliver? Do you have the date on that? How many months ago was this from Ollie trying to convince us to— this is from 3 years ago almost. Of Ali pitching us on his better way to hire engineers and evaluate them. Shout out to Ali. And whatever we invest in, we're always investing in— great hair, by the way, Ali, back then. Yeah.

I know, I lost.

Yeah, we all do. So we're all going in the same direction. But this is what I tell my new researchers, Ali, and it's the challenging founders who are odd or otherwise different that changed the world. A pitch from a Jenny's Ice Cream. Do you even know what flavor of Jenny's Ice Cream you were having back then? What were you getting?

That was actually outside my Starbucks office, which I was at for a couple of years actually. And then I moved to an actual office, which was first WeWork. And I realized that the sensory input I was having in that Starbucks was actually very stressful. Noises, the smells, and so forth. But I didn't realize it back then.

Imagine if I had this for Robinhood and Uber and DataStax and Thumbtack, the first pitches. Anyway, this has been a great episode. Thank you, Wade. Thank you, Mikey. Thank you, Ali. They're all hiring. So go to MicroOne. What's the— do you have a URL for hiring?

Yeah, microone.ai/research, specifically for researchers.

All right, we need researchers. So if it sounds great, you want to make a couple of hundy couple thousand dollars a month, tens of thousands of dollars a year, and you're an expert on something, reach out to my guy Ali. Mikey, if somebody is passionate about music and understands technology, where should they go to apply?

Oh man, this is embarrassing. I don't know the, the URL off the top of my head. It's probably suno.com/jobs, but just Google it.

Yeah, and your suno.com is your domain. And who knows, maybe the CEO has the proper email address. You never know.

I'm sure you can guess my my email address.

You might be able to guess it and send him something, uh, incredible. Wade, I don't know if you're hiring right now or you're busy, uh, automating all these jobs away, but I assume you're hiring some positions.

Yeah, yeah, sapero.com/jobs. But like we were talking about earlier, just build something interesting, notice a problem, and drop, drop us a line. Uh, I'd love to see, uh, how you can— what your ideas are for making the product better.

If you want to get a jump on the This Week AI launch, which will happen I think officially in February. This podcast is going to be in the This Week in Startups feed during 2026. So if you're listening to this, and you're a This Week in Startups listener, that's fine, you'll see it in your feed. However, you're going to want to sign up for thisweekinai.ai. And you can get our daily newsletter, we just bullet point the most important stories, and you can subscribe to the YouTube channel, all of that is available at thisweekinai.ai. We'll see you next time. Bye-bye. Great new hit song powered by Suno. What's this track that you made? Is this like— here we go. Look at this. Mikey, what do you think of DJ Oliver?

I'd go see him do this live for sure.

Really?

You might get a lounge gig out of this. Is there a lounge, Oliver, that you could spin at? You're in Austin. All right, coming up to you live from W— you compare it to the new sample. KCRW, the sound of Santa Monica, 92.3, easy listening. I did this, I did— I like used my voice and recorded a clip and then put it back into Remix, and I made a little guitar solo that was a little different. Oh, there you go. So it looks like it only really changed the first half, but pretty cool. Awesome.