ex-Google Director Just Revealed What's Coming Next...

Thank you so much for being here. We got a couple great things to talk about. Um, I just realized I don't have my notes. So, but yeah, all good. Yeah. Yeah. Yeah.

And as I was saying initially, the chance there's so much tech stuff that we're using here, from Riverside to OBS to a whole bunch of stuff, that the chances of this live stream going perfectly without any uh, tech issues, there's 0% chance of that. I'm just warning people. We're also all on, we're also all on ketamine, too, like Elon. So, I mean, you know, it's just the way it is. High. I also created a presentation for us later on, going over Scale AI and then all this acquisition craziness, but we'll love to hear what you want to go through, Wes and Joe, first.

Yeah, absolutely. Well, uh, yeah, Joe sent some excellent papers. Um, I was only, I only covered one of them before, so we're definitely want to kind of go through it, talk about those. Um, how interesting is, what is, is the stuff that Scale AI is doing with, with Meta? Is that something interesting or not really? Okay. All right.

And yeah, it definitely seems like there's more and more of an overlap between, kind, there's more like DoD and all the Tech Valley stuff and war and stuff. Just, yeah, definitely more and more overlap. So maybe um, we want to touch on that because you are, you guys are very much into, you guys are kind of like the insiders. So for a lot of the people on the live stream, I think this is going to be a real insider look, you know, and again, obviously your opinions are not that of Google or any companies that you've worked for. Obviously, this is your own stuff, right? Um, but would love to, to hear your opinion on, on all of that.

Um, we're confirmed conspiracy theorists. Exactly. We have our, I just took off my Star Trek uniform and was going to talk about, oh, you should have worn it for the show. I know. Just, I'm going to go old school like David Shapiro and talk about, are you, are you a security guy? Is your, is yours gold or red? Mine's red, dog. Red team, dog. I know some people are blue team and, and I just want to know who's going to make it through the episode. No, I'm red team. First to go, first to get killed. That's just the way it is, dog.

All right. Um, I'm not part of the Screen Actors Guild, unfortunately. I got to get killed early. Uh, so what you could do is I create a presentation real quick because I'm a corporate sellout. And uh, let me get that thing up real quick. Okay. So meanwhile, while you do that, I'm going to mess with my mic. So kind of ignore me. Okay, cool. And then Joe, Wes, feel free to just like cut me off anytime you want to bring something in.

Did the model pick this photo for you? Yeah, I, I, well, actually, take it back. I said Mark Zuckerberg and it came up with this, with his, with his chain. So, it's either stock or something. That is really impressive. Yeah, exactly. So, and then I have a live stream comment saying, "This sick guy blocked me from his channel when I talked bad about Sam Altman." No, I probably blocked you because you're an… Okay, so Meta x Scale AI, the 14 billion power play. Meta is finalizing a 14 billion cash for equity deal for Scale AI. Um, this is Facebook's second biggest M&A after WhatsApp. So when they bought WhatsApp, they p, they spent 10% of their market cap for that company. What was your reaction to the WhatsApp acquisition at the time it happened? Well, cuz I was a stupid American who never touched the glory which is WhatsApp and I'm like, what the hell is, what are they doing with this deal? It's like a, it's like a chat messenger. Why not using text message? And then I finally got tons of awesome um, Indian people moving into my house cuz I rented out like in Silicon Valley and they were like, if you want to talk to me, you're going to use WhatsApp. And then now I use WhatsApp for everything. And it just completely makes sense. And then finally, Joe, the news came out that WhatsApp is going to be monetized now. They're going to finally start putting ads into it.

Yeah, you've been advocating for that for a while. You got to monetize that, baby. I mean, it's the way it is. Um, and so, and I look at it as Zuck is basically saying, "Hey, you know what? If we're putting alms in all this, this inference cost is going to stack up on me and the pre-training cost and all the capex I spend, I got to figure out some way to justify this to the stream." So, so anyways, the Scale AI deal, people are like, oh, they're paying $14 billion for Alexander Wang, but if you look at as a percentage of market cap, it's less than 1%, so this thing could completely, and yes, I know there's like couple trillion worth now compared where they were, but this thing could implode, it's still like a rounding error for Facebook. I think it's a good risk, um, high upside potential, minimal downside. So, Alexander Wang is to lead the new super intelligence team that Mark Zuckerberg is recruiting. Um, and then he's also calling engineers and offering them the eight to nine figure salaries, just like, please join me. Um, and then also this is to basically get them back into the game after the catastrophe which was Llama 4. Like Llama 4 was just, oh my god. I mean, Llama 2, I literally was like showing rap videos and say that's Mark Zuckerberg because he has so much swag and he's so cool. Or as a kid say he's so lit or something. I don't know. I'm an old, old millennial. And then Llama 4 came out. It's like, what, what's going on internally? Like Joe, what, what happened? Why did that turn around so negatively and so quickly? It felt like they were trying to beat some internal, or not internal, but uh, public benchmarks and they maybe stretched a little bit, cheated a little bit. I don't know what the right word is. Uh, and then they got a very negative reaction because they advertised that, did that they did well on these benchmarks, but the model didn't seem to do well on anything else. And so I think people lost their enthusiasm.

Right. And then um, shout out to machine learning street talk guy. He basically did a 25-minute video saying like this is all, this is Goodhart's law or Campbell's law for social science nerds where you basically say, okay, we need to increase employment. Let's focus on that number. And then people will play games and say, "Okay, I'm going to hire people to like start digging ditches with spoons." And so it's all the same thing here with these benchmarks. Um, but people are probably wondering like, you know, what's going on in M&A? This is all confusing. Jordan like shed some light on it. And I just got back from Holiday Express, so I'm getting my best take at this. You got three major M&A deals: Aqua Hire, license and release, and stock purchase. Now, Aqua Hire is basically a glorified hiring exercise. So, anytime you hear any of your homies or whatnot try to like uh, shine and say, "Hey, well, I got acquired by Google or Facebook," you should ask them, "Was it an Aqua Hire deal or a full stock purchase deal?" And if they say Aqua Hire deal, that basically means they didn't get really paid anything, and Google said, "You have no enterprise value. We don't want your company. We just want to hire all your employees out, and then we might give you some additional retention money, which is golden handcuffs." I don't know if anyone here has watched Silicon Valley and HBO. There's a Joe won't watch it because it's like the uh, the story of, of Dorian Gray where if he watches PTSD, it gives him PTSD. I watched one episode and they did an M&A negotiation and I was getting PTSD from it. Um, but there was a scene where all the founders are on top of the roof and they're just playing video games and they're like, "Wait, what are you doing here at Hooli? They, they acquired you and you're not doing anything." It's like, "Oh, we just rest in vest. We stay here for three years until our retention pays out and then we leave." In the meantime, corporate doesn't want to do anything with us. And so here that's where we are. And when I was an admin with Joe back in the day, one of our directors I was also an admin for, he got acquired. And for some odd reason, the Google product team didn't want to do anything with him even though he was like a genius. And so he just sat in his office and invested and he would go mentor kids and things like that and he would raise his hand every time he'd be like, sometimes the VP saying, "I know what you're trying to do but probably won't work." And they would like ignore him and then the thing would fail anyways cuz he didn't listen to him. So anyways, most of the deals back in the 2010s were Aqua Hire deals because most startup ideas fail. And then you have these license and release deals where basically it says, "Okay, we want your team and we also want some of your IP cuz your IP is decent, but we actually don't want your business because maybe you made hot dog not a hot dog app, which is just stupid, but it had a nice uh, image recognition um, uh, uh model under it that we're going to try to reapply to something else." So, back in the day, those license and release deals, you maybe got um, some money back on your dollar invested. So, a VC in an acquire deal gets nothing. But for license and release, they might get 85 cents a dollar um, if you're lucky. Maybe you might get maybe 10 cent, maybe you might get your full dollar back or something. But then things change in the regulatory environment. We'll go into that later. And now we're seeing all these AI companies saying, "Hey, license is back in vogue. So, we're going to start doing those deals to get away from uh, regulatory heat, which I'll go into in next slide. Then you have full stock purchase deal where everyone gets paid uh, when Salesforce acquired Slack. Um, when Google buys Wiz, that basically means we're going to pay a gigantic premium on your equity. So, an investor is going to get maybe a 100x or a 1,000x or something ridiculous. And those are the big awesome deals where people get rich. Um, so, let's go to the next slide here real quick. And then Wes, if you have any questions or you want to stop me, just let me know because I'm one of those talking heads majored in political science. I won't stop talking to you. Give me money. This is great.

So Joe, I'm, I'm sorry if I did, I cut you off during your intro. I'm sorry if I did. No, no. Okay, I'm good. Okay, sounds good. Yeah, a gradient check says Joe could use some color in his background. Yeah, Joe, why don't you have some color in your background? Some pink, some blue, you know, a nice painting. Actually, Joe has a Japanese watercolor that the uh, Tokyo office gave to him because he's so awesome. It shows Google and then one O is a rising sun and the other O is a Japanese version of Joe watercolored. And I remember animated. I before he became my admin that was sent to me before I met him and I saw this I was like, who the hell is this guy? Who does this guy think he is, stealing this guy? This guy's an, you know, and I met him. I'm like, ah, god damn. Now I'm gonna make a watercolor for him. Okay. So why, and Joe feel free to cut me off whenever if you see something you want to add in or Wes, um, why licensing deals fly under FTC radar, um, a lot of things for FTC's there, it's based on the antitrust laws that were used to bake about break up the railroads because back then, time uh, the railroads, oil companies in the guild age, the people thought they had too much power and they're using that influence to squeeze small businesses and whatnot and so what the FTC looks at is if there is a uh, a market of, let's say search for instance, we all know that um, if Google went to go acquire another search player like Bing or something which would be wouldn't make any sense, the FTC say, hold on, Google, you already have like 80, 90% of the market, you then bought Bing and you, let's say you got a few percentage points more, the FTC would say, oh well, you kind of already do have a monopoly and working on that, but now this is like a super monopoly so we're going to prevent that from happening because the theory is you're going to raise prices or something. Um, and then, um, but when you do a license and release deal, you're not buying the company. You're just taking maybe some of the IP and the headcount, but that organization still exists. And it can go off and die for all you care. But what's important for you is you're not going to get the FTC review like it used to do. Now, the FTC could change and say, well, in the Scale AI deal, yes, Meta owns 49%, doesn't have a majority, but for all intents and purposes, Alexander Wang is still on the board. You're probably going to get board seats. Alexander Wang is going to get even more money to stay at Facebook probably in stock. So, do you think he's going to be aligned with what Facebook wants for the future of Scale AI's company? Yes. So, they could say it's a de facto acquisition even though it's not 49% and they could investigate, but we'll see. Mhm. Um, then when you do these type of deals, you get regulatory fasttrack. So, I don't know if you all saw three days ago, um, Google announced they bought Wiz months ago for $32 billion, which was that deal happened. They started that company like three years ago, and then they get a $32 billion payout. And that was their second acquisition because they got acquired by Microsoft back in the day. So, now their only issue they're dealing with now is they want to get individual, not just yachts. They want to get yacht craft carriers so they can have supporting yachts and helicopters coming in. But the issue with that deal is news just came in um, uh, that uh, I should read the live stream comments while I'm talking. News just came in the FTC is doing a review on Wiz. So now they have to go through a year-long approval. And the FTC does not like tech right now. Uh, the Republicans hate them, Democrats hate them. And just like the Figma Adobe deal, you could see things fall apart. Now, the Adobe deal fell apart because the UK's um, CMA, which is their version of the FTC, was going to block that deal. And then so, uh, Adobe said, you know, we're just going to walk away and pay a gigantic break fee. I think it was maybe hundreds of millions, if not billions of dollars. Oh, it was a billion dollars. Billion dollars just to say, you know what, we actually paid too much for this deal. Here's your billion dollars. And you probably actually did wildly overpay and their stockholders were very happy when the deal fell apart. Joe, go into that because that's you worked at Adobe back in the day and you know that industry.

Well, it's amazing that Adobe wasn't able to compete with Figma. I mean, they had a huge amount of warning. You know, the apps were moving online. They were becoming collaborative. You saw it in all the productivity apps and Figma's first product was a drawing product which really would have competed with Illustrator uh, on the Adobe side. and it didn't really take off in the way that the Figma founders wanted. You know, they wanted a much bigger business. So, they went back to the drawing board and came up with a more design focused product that became the Figma that we see today. So, Adobe had at least two years of warning that this thing was coming. They spent a bunch of money and built their own team and probably over five years tried to build a competing product. I forget the name of it. It was something MX, but it, it never uh, managed to generate much enthusiasm. They eventually shut it down when they did the Figma acquisition. Uh, and then when they acquired Figma, they paid a tremendous amount of money. I mean, Adobe is significantly smaller than Facebook, so I'm guessing it was several percent of their, of their market cap. And I think the amount they, they paid was just below the threshold where the board of directors actually had to vote for it. So like the CEO and his team sort of put together this deal that would just get under the limit and they didn't need approval because I'm pretty sure they would not have gotten that approval. Uh, and then when the deal fell apart, everyone kind of rejoiced and their stock recovered, right? No. No. Well said. Yeah, they definitely were overpaying that one. And a good sign of if you're overpaying for a deal is what, how badly Wall Street will hammer the acquiring company stock. You'll see it usually get crushed if it's way too, it's way too much. Yeah. So, just as a an exclamation point on it, uh, Meta stock is up the last two days. Good point. Significantly, like people are rejoicing that Zuckerberg is doubling down on this AI goal, right? Two good pieces of news. One, he's restructuring his AI or, and two, he's monetizing WhatsApp. Um, I know everyone here is like, "Oh, you corporate sellouts and blah blah blah blah." We're just telling you what Wall Street thinks about what's going on here. Just all we care about is if you like and subscribe Wes's channel and if you want to, our channel, too. So, second, the most sellout. I am a corporate sellout. Look at me in my meth lab over here. Uh, so the other thing about license and release people don't talk about is no liability from acquired company. When you acquire a company, you acquire its whole entire uh, legal history and its liability and from they could have done some effed up things before you acquired it. You still are going to be on the hook. And now lawyers are going to say, "Well, it was a company worth $und00 million. Probably don't want to go after it." Well, now it's owned by Google. It's multi-trillion. Hm. This could be interesting for a class action lawsuit. As a side note, one acquisition I worked at Google, it was one of our quickest closed deals. We clo, we got notified of the deal and we closed it within like two to three weeks and it was a, it was a Aqua Hire and it was from company was called Homejoy and basically what they were doing is they were going to be the Uber but for cleaning services at your house and this is before uh, Uber went to war state of California and got the proposition to classify um, uh, uh Uber drivers as contractors and not full-time employees. Mhm. uh, Homejoy wasn't there yet and so they were doing well in their business but did not have uh, as big as pockets as uh, Uber did and so they were right in the middle of raising their next round of funding and then the state of California raised a lawsuit against them and so the woman who, who was organizing, I forgot her name but she was, she was awesome, she then effectively was like, okay, well, I guess we're not going to have any funding because all of our VCs pulled out so then she came to us and we were able to do an Aqua Hire where we're like, "Hey, we want nothing to do with the state of California lawsuits, but we'll make sure all of your employees land here at Google and get good jobs." And so, we closed that one pretty quick. And it was a one of my favorite deals because we gave all those people good jobs and we didn't give them pink slips and they were able to pay their mortgages and do great things at Google. So, let's go to the next slide. License deals are cool again like I mentioned. And so, here are the major ones. We have Microsoft inflection license release deal. That was a $650 uh, L&R. Um, and that kind, that was one of them. And then we have Google and Character AI. That was for $2.7 billion. That's where Google got Noam Shazeer and a few others. Noam was one of the authors on attention is all you need and just really, really good engineer. Got about 30 researchers from that. And then Amazon did the Adept deal for $330 million. And then we had the Meta, the Meta uh, Scale AI deal which is an investment but it feels like a license release because they're also um, getting the CEO plus a handful of employees, right? Um, so let's go to this, Metascale AI deal feels a lot like the Google Character AI deal. Yep.

Why don't you go into that for a second? I mean, as you said, they're sort of extracting the CEO, founder, and a couple of key researchers. That part's very similar. Uh, they're structuring it in a way that sort of avoids regulators. That feels the same. Mhm. Uh, and it's an incredible amount of money for what amount, what looks like an aqua hire after you sort of look at all the components of the deal.

Exactly. Um, but now, uh, we'll definitely have to go into the valuation point you mentioned; have a slide on that, but you're right, it looks like an incredible amount of money. So Scale, founded in 2016 by Alexander Wang and Lucy Guo, um, they do data labeling and evaluation, and they have two different sub-organizations. Is Lucy Guo staying with Scale, or is she coming to Meta?

It looked like she was staying with Scale.

Interesting. Yeah. I didn't because she's pretty prominent, so they would have mentioned it. Oh yeah, and also my co-founder Lucy's joining us because he sent out a note and he just mentioned, "Hey, I'm still going to be on the board of directors of Scale," um, and you know, very little is going to change. We're going to get our chief of staff to lead; he's going to be the CEO; he's chief of staff for strategy; that person's going to be the CEO, but nothing Lucy. So they have two sub-suborgs: Scale AI one, or is basically they do PhD-level data training sets that these LLM providers pay for, like OpenAI and Google. Used to be Google. Google's now signaling, because of the deal, they're backing out. Have another. So is everybody else too, right? Everybody's bailing out from using them, right? I, this is breaking. Did you hear any other companies?

I think so. Yeah. I forget who, but like all the big players, I think, signal that they're getting out. They're not using Scale. OpenAI. From what I heard, OpenAI says, "We're cool with it. We're going to stay." Okay. It'll be interesting to hear what Amazon's doing or will become a major player. They've been kind of sleepy over there, but you know, um, and then you have the other business, which is more of like Amazon Mechanical Turk, where it's like 100,000 employees and they can help you, uh, contractors and whatnot, and they can help you with data labeling for other types of data sets and whatnot. Um, and then so let's go to the next slide.

So what is Meta really buying? I just look at this as getting Alexander and energy into the organization so they can hopefully turn things around because it just became kind of a cluster long before. Um, and then they're going to have the ability to use Scale AI's data pipeline to help them train models. But the most interesting thing is me and Joe have been talking on the show for the last year and a half about synthetic data and saying, okay, there's more and more synthetic data coming out. There is now RL with verified rewards. There's now models coming out that you can give it a confidence level and it can say, you know, "I'm going to trust my gut," and it can actually increase its performance to a degree. Like this has to have downstream effects on Scale AI's business. And so we saw some of that in regards to Scale AI missing its revenue target last year. It was supposed to hit a billion dollars instead it hit 870 million. I mean, to us mortals, it's still a lot, but it wasn't the growth as fast as they expected.

Um, and so they were getting some pressure from that. And then also, if you look at the valuation, the amount of money that Facebook put into this deal compared to their previous valuations, it wasn't such a big of an uplift in exit value, which made me think of maybe they were realizing we might get to a point where we're getting to the end of the line of this type of business. Um, Joe, you made some good points during our conversation about the enterprise sales cycle, and maybe they were seeing something from that too. Maybe you can go into that.

Yeah, I think there's, you know, a lot of ways to signal that deals are delayed. Uh, and so if people start to think, okay, maybe I don't need this data set, or maybe there's other ways to generate this data, or maybe synthetic data is going to become available, or I can produce my own synthetic data, or any of those things are true. A typical thing in enterprise sales is that you'll just see the sales cycle get stretched out. So signing the deal will somehow just get delayed. And then salespeople are looking at their set of ongoing deals. And if they see that they're all tending to get delayed longer and longer, that's a really bad sign, right? And if your book is getting delayed that way, you kind of assume that things are slowing down; many of those deals won't actually go through, and that your sales cycle is somehow being impacted by something in the environment, and that's a sign that either the product's not right, or the economic environment is slowing down, or something's going on that's big. And if that was happening to sales at Scale AI, I could easily imagine the founders getting nervous.

Exactly. And so you're probably thinking, okay, maybe it's time now to get out while the getting's good; you know, we can't take this anywhere higher, right? Maybe this is the peak, and it's the right time to do a deal.

Exactly. Gotcha. I look at this as a mega ultra super success for Alexander, getting to this point. Um, I see a lot of hate on Twitter saying, you know, "Oh, he has no value, and all they do is contractors" or something. I'm like, "Oh, okay, go get yourself a 154 billion dollar exit. Uh oh, you can't then go back to Wendy's." So I think everyone should be applauding what happened here and not stressing Facebook so much on, "Oh, they spent 14 billion on him," when it's like, "Hey, they have this capital, they need to reinvigorate their AI, or if they're able to get that show on the road and get things going, then that's a huge opportunity for them to increase revenue if it's done right."

Um, so let me go to have one more slide here. Um, so kind of like risks on the horizon is just Google walking away from the Scale AI contract, which I don't think Facebook really cares about because they're more focused on getting Alexander and crew. Um, but other customers could leave. And then there's a question of are they overpaying for the price? The multiple is about 12 to 15x on their revenue. We don't know if the rest of this year if the revenue could hit a billion or be less than that. And then the other thing is just cultural mismatch. Alexander is 28, works at a small company, knows how to move things and ship product and get stuff done. He then goes to Facebook, which is a gigantic government and works through politics; it's very hard to get stuff done. He might have an alignment with Mark, but then getting what he needs done through the rest of the organization could be like pushing a safe through sand. So it's yet to be seen how that's going to work out. Joe, you used to work at Facebook. Any thoughts on potential stumbling blocks Alex might deal with going into that organization?

Well, I think there's always risks in a big organization like that. You have people who are already there and sort of consider that this is part of their charter. I mean, Facebook already has a couple of teams working on AI and ML. Um, there was just a release out of Yan LeCun's team just I think day before yesterday. Um, so that's probably the biggest issue. And then, you know, from his perspective, is there going to be internal movement? Like are people from other teams going to join his team, or are people going to try and recruit away his best people? So I'm sure there's jockeying for position going on. And then lastly, Facebook itself is still under investigation by the regulators, right? So it's a kind of uncertain environment for them, right?

Has everyone here seen the movie Office Space or Half Baked by any chance?

Mhm. And there's that scene where the guy is sitting down in prison, and then Squirrel Master comes up and goes, "Hey, he's mine. Like, don't mess with him, cuz the prisoners are trying to attack him." Yeah, Joe, you had. Yeah, I don't know if you can. Well, put you on the spot. You had a similar experience on your onboarding at Facebook where you're getting like recruited. Maybe you can tell a little bit about how they have this weird way of allocating engineers to certain teams at Facebook.

Yeah, I think that's actually one of their strengths. Um, they have a really interesting training program they call the boot camp. And so new hires coming into Facebook are not normally allocated to any specific job or team. They just come in, they go to this boot camp training, they are there for anywhere from a couple of weeks to a couple of months, and they go through training sessions where they learn how to do things inside the company, including checking in changes to some of the products very quickly. So usually they have a goal of, like, first day you check in a change to one of the products, and then on the flip side, all of the teams inside Facebook package up small changes or bug fixes that they want, and they've sort of curated them in a way that new hires coming in can pull something off the queue and then go through it as an initial project. And then there's someone in that team who helped file that bug or that issue is listed on it, and if the person who's coming in as a new hire and working on that bug has questions, they can contact that person. Well, this is a perfect moment for the existing employee to sort of see if the new hire is someone they want on their team, right? And they're doing it in an environment where it's, you know, low stress, well, at least low stress for the existing person, uh, and a chance to evaluate this new person. How good are they? How fast are they? And then if they see a person they like and that person's showing a certain facility, they'll try and recruit them on the spot, right? And ideally for Facebook, the new people they're bringing in are getting recruited by two or three internal teams in their first couple of days, right? Because that means they hired a great person, and it means they're showing a lot of skill, and it means the teams have a constant stream of new hires coming in. But it's weird for most people because they've separated the hiring and recruiting from the placement, right? You know, the assignment to the teams.

Now, tell the funny story about how they were fighting over you at lunch.

Well, when I when I went there, I already knew that I was going to build a team to work on privacy. And so I was, you know, in a so-called allocated person. I already knew what team I was joining. Uh, but I still wanted to go through boot camp because I thought it was an interesting possibility to, like, learn all their systems and also to see how their recruiting program worked. So I told them when I joined, you know, "Put me in the boot camp like normal," uh, even though I was going in as a manager and not an IC. And as I was doing, uh, you know, bug fixes one by one trying to hit all the different systems that I was curious about, people would say, "Oh, you know about this stuff." Like a couple of the bugs were around doing JavaScript on the client side. So I knew about that. And another set of bugs were around localization, and I had a long history with that from much earlier. And then a third set of bugs were around advertising, which I knew a little bit about from Google days. And so each time I would be fixing a bug, I would talk to someone who was listed and I didn't really understand their system yet, and I would engage with them and ask all these questions about, you know, "What's the issue? What do what do you think of these possible fixes? How should I pursue it?" And inevitably, they would say, "Oh, you seem like you're interested in this. Would you like to join the team that works in this full-time?" And then I would say, "Oh, I already know which team I'm joining." And then they would be sort of frustrated like, "Well, then why are you still fixing bugs and still in boot camp?" So I was kind of, you know, going upstream at that point. There's a one part where I guess he was sitting down to eat lunch and two managers kind of pounced on like, "Hey, so it looks like a team." And then someone super senior like Scroll Master came through and was like, "Nope, he's with me. Let's go, Joe." And both the managers kind of like scurried away.

I think that's a, it's a core competence at Facebook if you're a team leader that you are aggressive about hiring because they are in fierce competition, and if you're really good at it, you are there in person; you come to the area where people who are doing the boot camp are usually sitting, and you sort of mix with the teams and you try to put faces to the names that you know from interacting in the database.

Yeah, that's an interesting thing. Yeah, how that's set up. That sounds like it would work very, very well. And I like the idea that you don't get just somebody dumped in your team. You kind of have a chance to kind of see what they're like working and then request that they get pulled to your team. It's kind of an almost like a second round of, I mean, it's placement, but it's a second round of recruitment almost like you you make it to the inside and then you're recruited onto the team. That's an interesting approach. Is that common or is that?

No, it's very uncommon. And you know, it's a fairly expensive thing for Facebook to do, right? Because it means all these people are delayed from officially having their job and their assigned tasks for, you know, up to a month while they're in this boot camp. Uh, and I and you know, they pay all these people to be trainers. So there's like people walking around, and then people from the various teams give classes like hour-long sessions on how the different systems at Facebook work and how to do various kinds of work at the company. So that's also a sort of expense for them. And then there's this huge area where all the new people are attending classes and you know, going through their projects and getting help from sort of hall monitors and so on, and that's also expensive. And then all the teams are carefully curating bug reports and improvement requests and putting them in the database knowing that they, you know, it's not it's not like they're setting aside work for themselves. They're purposefully setting aside chunks that can be done by a new person, right?

Yeah. So they have to carefully make sure all the information is in there and it's complete for a person with no context. All of that is very expensive. Uh, and I credit Facebook for putting in the investment. A new person coming in has a tremendous opportunity to learn all about the systems, all about the teams, how the company operates, how things get done, uh, literally make changes to a running product on your first day, which is extremely dangerous. Uh, they're very proud of the fact that people, you know, new people coming in bring down Facebook every couple of months, right, by accidentally checking in some bug. And then lastly, it's a great way for all the teams to get direct exposure to these new people and vice versa.

Yeah. And then eventually find their match. The try before you buy it. It's great. And then also speaking about introducing a bug, I was listening to some podcast, and the director, she was given an assignment, and she actually brought down part of Facebook's site to a certain geography because she did somehow do a denial-of-service attack on Facebook, and she was petrified, and the engineers came up and, "No, no, you found a vulnerability here that we didn't, you know, think of; now we're like we're going to fix it." And now to compare this to how Google does things. I used to go to Google's new noodle orientation all the time because when we'd acquire a company, I was the head head HR contact, and my job was to prevent the rest of HR from hurting my people. And so like these are mine. And so what you would do for Google's orientation is you would sit like a week, week and a half long and just in just meetings and videos about like our orientation and what we're doing and things like that. And you'd very gradually maybe get exposed to the codebase, whereas day one at Facebook, like Joe was like in there like doing things, and a lot of people appreciate that having the ability to just jump into the into the code stack and get to work. So so yes, um, yeah, what else shall we talk about?

Well, so I do want to touch a little bit on all the incredible research that's been going on. So I mean, briefly, I feel like this is we don't want to beat a dead horse, but how about the Apple paper? Let's. Can we briefly just talk about that? What do you What do you guys think about that? You want to start on the controversial stuff?

Yeah. Yeah, Joe, go first because I gave a whole class the first 40 minutes. Your turn. Your turn.

I'm not sure what's motivating Apple. I mean, it's strange because everyone, well, investors perceive that they are going slow on AI, and it's not clear that that's because they don't think AI is ready for their products or they don't have their own story together internally, or what exactly the problem is, but they're definitely lagging other large tech companies, and this is the at least the second paper that I've seen where they mostly spend their energy pointing out how these AI systems are not good enough, not ready for prime time, and then people sort of react strongly at first like, "Oh my god, you know, these AI systems can't do something important," and then gradually sort of realize you can pretty much do these things if you just use the model a little better. Yeah. And then the paper kind of disappears into irrelevance. And I feel like that's happened at least these two times. So I'm not really clear on what Apple's thinking with this kind of research. Do you do you agree with this, Roth? Is this is this sort of the direction you perceive?

Yeah. I mean, yeah. I don't understand what the point behind it is. I um, I've seen a few number papers like this, or some of them were blog posts, some were papers, like one was literally called "LLMs Can't Reason." And yeah, there's so many things I don't even know where to start because number one, you know, if we're talking about if something can do something or not, like can humans run a 4-minute mile, right? Um, now if we see an example of a thousand humans not failing to run a 4-minute mile, does that prove anything? No. I mean, you can strongly suggest that maybe it's impossible, but you need one example of a person running a 4-minute mile to say, okay, so that's disproven. Um, and then so one, you can show me a million ways LLMs fail at doing something. Does that prove that they can't reason or can't think or whatever? That's one. Number two is what do we even mean by think or reason or anything like that when it comes to LLMs? Because that's kind of a very human-centric thing. So it's like if you ask these people that say LLMs can't reason, okay, what's an example that would disprove that hypothesis? Give me an example, like if I get them to do something, right? What would prove to you that they can reason? If you can't come up with an example, then this is like this conversation doesn't make sense. Um, also in the paper, they have the river crossing problem, which is impossible for after, you know, N five plus or whatever, like basically at that point it's impossible to solve. So the model probably says it's impossible to solve, and they mark it as zero. Um, so it's just there's so many things there also. Why is, you know, being smart enough to realize if a model is smart enough to realize I can't do this problem within my context window, but I can create a tool that solves the problem, then builds that tool using Python code or whatever and then solves the problem, like how's that not reasoning? Like why are we it's it's strange like why we choose that.

As the definition, um, and the previous LM can't reason paper from a couple of years ago was not of Apple or somebody else, but usually what they try to do is just hit it at some limitation that the model has.

So before, a lot of the puzzles would be like, you know, um, it can't count the number of words in a sentence that it's about to say, right? Because it didn't have reasoning yet. So it couldn't think through it, get that data, and then count it. It had to, which humans can't either. Like I can't predict the number of words that my next sentence will have before I say it unless I write, you know, I say it first, count it, right? So a lot of the problems were like, yeah, like furniture placement where you had to meet certain constraints. Now that we have reasoning models, of course, it will 100% succeed at all of those, right? So now there's strictly, and the other paper also had, you know, the context window limitations, so just hammering it in places where the context window, uh, would fail, and this Apple paper is now 100% context window. So yeah, it's it's just one of those things. It's like so many faults with it. Like number one, what is reasoning? Number two, what would be an example of an LM doing something that would qualify as reasoning? At which point you would say yes, they can do it. And then yeah, three, don't just hammer their existing limitations. Um, that seems weird. Yeah, I'd be much more excited if someone like extended the capabilities. If they said, "Here's a limitation we bumped into, and here's the things we did to try and overcome it, and maybe one of them succeeded." That would make me much more excited. Like that's a contribution. Uh, I like your question. You know, what would have convinced you that the models are able to reason? That's a great question.

And then lastly, I would say to them, when you publish a paper like this where you say, uh, generically models can't do X, there's a real danger that the reaction is no, you can't get the model to do X. And someone turned around immediately and got, uh, one of the models, I think it was 03 Pro, to do the Towers of Hanoi with 10, uh, discs, which was one of the more difficult problems in the paper, and it did it correctly without tool usage, which is kind of amazing because that's a long sequence of moves. Um, and then someone else did it with tool usage, I think, on an earlier version of Claude. So, like one of the major examples in the paper was already disproven by some random person in the community within a couple of weeks of the paper being published. Like, that's kind of sad. It means you didn't really put out much effort to prove your core thesis. Yeah. Yeah. Yeah. There's a lot of problems. Yeah. I I I agree. Plus one to what we both said. Old Google+ parliament. Sorry. Heart. Old habits die hard.

When I think I see all of these different research papers attacking these LLMs or people going on the press tour like Yan Lun and attacking them and things like that, it reminds me of this quote from Max Plank. He says, "A new scientific truth does not triumph by convincing its opponents and making them see the light, but rather because its opponents eventually die and a new generation grows up that's familiar with it." That's good. And I look at my my niece right now, and she gets to play with ChatGPT and talk to it, and she's happier than I'll get up. And I wonder what their generation is going to think 20, 30 years from now if they'll be so focused on these questions like that. And that's why for myself, benchmark and benchmark porn, I mean in the initial days when GPT-4 came out and whatnot, seeing the benchmark leaps and how things improved, I like, this is cool. This makes sense. Then as time went on, the games other companies were playing by doing like, okay, we're going to compare one shot with GPT-4 versus 10,000 shot with our model. Look how we've improved on this benchmark. Wow. And so I started focusing more on things like Swed Lancer or other measures of, you know, Fiverr, for instance, when ChatGPT got released, their job postings went down 17%. Because people were saying, hey, I could use this model now, and I don't need clip art or copy editors, or Stack Overflow is seeing their traffic completely implode. Um, for me, the big indicator is going to be more benchmark freelancer where it's people are exchanging money not knowing they're exchanging money to an LLM because they think they're working with a real developer for a fixed problem. Now to be clear, developers, I think there is a lot of job security for y'all in the future. Engineering work or so much more than just coding. It's just setting up problems, dealing with internal or technical complexities and things like that. I'm not one of those who thinks all your jobs are going away. Um, I like those type of indicators more. The big indicator for me is when am I going to hear my friends who are working at all these different tech companies say, "Hey, we decided to forego headcount because I want an AI agent instead to take on a specific role." Have not heard that yet. And that is something big for me. So, when I hear all these startups saying, "Oh, we're agentic this and that and this," it's like, "Yeah, you're using an LLM to augment people, which is cool, but you're not replacing headcount with that. Um, so anyways, that's my kind of tangent.

Oh, last thing to bring it all home. Do any of you remember what happened to Apple in 2012 by any chance, what the controversy was then? Okay, so Apple was like, you know what, f Google and Google Maps, f them. We're going to do our own. Oh, create Apple Maps. Apple Maps. And Apple Maps was like leading people into wrong directions. I organized an offsite for Google. We were going to go into the Santa Cruz Mountains. And I went to BevMo and got two shopping carts full of like high-quality booze. Then I was going to have Armadillo Willies show up with a van full of ribs, brisket, and everything. And then we had a park team that was going to show up and do geocaching. So, and you're in Santa Cruz, and the summer is beautiful. Our driver was taking us there in the bus. And I was just looking at the route he was going, and I was like, I think we double back somewhere. And I went up to him, and and I was like, hey, you okay? And he's like, oh, I'm sorry. This Apple Maps thing is putting me in the wrong direction. And everyone in the in the seats, they started laughing. No, no, no. It's like download Google Maps, it's going to be okay, or get to an Android. And that controversy was so big that the head, one of the senior VPs at Apple had to resign, and Tim Cook had a huge just sore spot from that cuz he's the operations guy. He wants everything to be perfect, fit and finish. And he probably never heard, he probably for that whole entire year or two years was constantly hearing people at him how they got lost from maps. And I think scarred into his brain was never effing again. If you're going, we're going to launch something on the iPhone, it's got to be perfect. And I think for the LLM side, they weren't able to get it to that fit and finish level, and they said, "F it." That's why Siri intelligence is not going to be upgraded until like mid-2026, which is kind of a huge, like I don't even know if they're going to hit that number. So instead, if you can't beat the other companies, what you do is said criticize them and start releasing your AI research papers and saying, "Oh, this technology sucks." Right? That's where Apple is right now. They can't they can't do so criticizing. Yeah, that's where I that's my read, but I'd love to hear, Joe Wes, what your thoughts are. Well, um, it's all yours, Wes. I mean, I again, I have no idea. That that certainly seems reasonable. Um, let's just I'm realizing we're this is really good. We're running a little bit late. Let's hit a few of the more fascinating papers. Yeah, because the intuition paper, intuittor, right? The learning to reason without external rewards. Let me click over here really fast. Um, well, actually, you know what I can do is I can just bring it here into the window. So, I'll block kind of our beautiful faces, uh, or somebody in chat put our our big beautiful, uh, smart what what do they say? Big beautiful bald faces or heads or something like that. Whatever that was though. Yeah. Thank you. Um, they call me they call me Megamind. So Megamind. Yes. Uh, yeah, that's good. Um, so learning to reason without external rewards out of um, Berkeley, and this is weird. So, I plan to do a video on it. So, for people that haven't been um, that haven't read this yet, it's strange. It's weird. I I don't 100% understand why this would be like intuitively doesn't really, uh, make much sense because it seems like instead of using some sort of verifiable outside external rewards, they try to look at how confident a model is in its abilities to answer a question. Um, and of course, the more confident it is, that correlates to it getting the right answer more, or if it's not confident then it's and confidence I mean they describe the kind of the math behind what they mean by confidence, um, is basically it seems like how many different sort of branching ideas it might have about how to answer it where it's a little bit more narrow that suggests confidence; if it's like it could be a million different things then it's not confident, but I guess they asked what if we train it and the reward the um reinforcement learning reward was it getting more confident on the answer, and somehow that improved its um accuracy. So, does that make any intuitive sense at all? What do you guys think about that?

Well, you're you're using the internal confidence as, uh, reinforcement learning kind of scoring mechanism, right? So, the confidence doesn't really, uh, tell you how to train the model. It just says this response was more likely correct, and then you can choose from the correct answers and the original questions to do RL training on another round of the model. Right? So there's sort of there's sort of two steps happening there. I agree with you though. It it does seem like you're getting something from nothing to use confidence to decide when the answers are more likely to be correct. You're like, you know, shouldn't I just know if the answer's correct or not? Uh, but before we saw this confidence-based mechanism, we saw things like self-consistency where you would just sample the model, I don't know, 16 or 32 or 64 times and then take the answer that was the most common, right? Which is also kind of strange. It's like how come the well first of all, how come the answers are different if you just ask the same question over and over again? That's down to the sort of statistical nature of the model. And then second, okay, if if there's a statistical nature, it's tending towards giving me the right answer more often, and the right answers tend to cluster together, whereas wrong answers tend to be more spread out, like they're wrong in different ways. That's kind of weird. Okay. And then I can use self-consistency to sort of pick out the answer that's more likely to be correct. Anyways, however I do it, self-consistency or this, uh, internal confidence, the end result is that I get the correct answers, and then I take the correct answers plus the original questions and I do a round of RL. The model that I get out of that is stronger than the one I started with, which is also a little bit iffy. And then I can repeat that process because now I have a stronger model, and it's even more confident about even more correct answers. Right.

How much of this, Joe, is do you think is like pruning in a way of we do pre-training on these models, and all of these techniques on the RL side is like a nice gardener of trying to cut away the noise and helping the model get to a point where we can get the the distilled information that we need so it improves performance, and let me know how terrible that analogy was.

Okay, that's a great question, and I think pruning is a good analogy because you could you could look at the what they call the pre-training, the sort of ordinary training on a large corpus of text data, and you could say that is giving the model lots of information about the world and many ideas, good or bad ideas, who knows, like it's it's internet data, so all kinds of different ideas, and the model is looking for patterns in that data. And we have to assume that correct answers are guided by that pattern formation, right? The patterns tend towards correctness, uh, because that's also again related to consistency. And so what's in the model at the end of that pre-training is a bunch of good ideas, but the model is still stochastic, and there's also bad ideas in that collection. And so your analogy of pruning is like identifying the correct, uh, reasoning traces or the correct directions and emphasizing them, which is what reinforcement learning is all about, right? The reinforcement part. You're not eliminating the bad answers. You're just, uh, adding weight to the correct answers and letting the bad bad answers kind of just fade. Mhm. Yeah. What kind of a body blow would that be if somehow they were able to figure out down the road, okay, we've located certain areas on the internet where people's ideas were just so bad it makes our models dumb. Anyways, I think there's a huge amount of, uh, data curation going on. Some of the papers in our in our recent list were about just cleaning up the data set, uh, and one of the ways they clean up the data set is by asking another model or, you know, ideally a stronger model to sort of look at the data and and trying to eliminate the lower quality information from the data set, and that also improves the training quite a bit. They were doing that with Cohere, weren't they? Cohere did a really nice job of that. They had a very big pipeline, and there was a recent, uh, paper I want to say OpenAI where they created a fairly large, um, open-source kind of data set using the same approach. A lot of filtering, a lot of quality checks. Nice. Yeah. And one of the papers we looked at it's also training, uh, changing the weights of the models based on kind of the synthetic data that it generates. We should probably look at that that next. But what we were talking about like pruning with RL. I mean, that's kind of what, uh, so Dwaresh Patel had several Anthropic researchers on his, um, channel where they kind of talked about this a little bit, like one of the researchers said, okay, like what if all of the whatever abilities are locked in the models, like once it's trained it's in there somewhere in the latent space, and RL, I mean instead of saying pruning he said it kind of like lifts out the the needed stuff. It's kind of the same analogy. Um, so, but that's kind of what they were talking about. So, it sounds like, yeah, it there's a lot more there than maybe meets the eye at first, but with reinforcement learning, we're kind of letting it emerge the proper things cuz yeah, it sounds like it has the data for or some signal for what's wrong, what's right. It just then the reinforcement learning kind of like helps it, uh, strengthen that signal, which is I I it's just interesting because it intuitively seems like it doesn't make sense. Um, but obviously it this would reinforce the idea that it's already somehow baked in there and we just need to kind of like get it out. It has a feeling of perpetual motion. Like it doesn't seem like you should get something with this approach. And then going back to the Scale AI conversation, I mean their business as far as I understand is producing large high-quality data sets, and a lot of those data sets are used for this kind of reinforcement learning training. And this list of papers that we've been recently discussing is all about either creating data sets from scratch or understanding how to reinforce the model without a bunch of hand-curated data along the lines of synthetic data or or internal consistency or confidence or whichever technique you you like. But that sort of implies that there's less value in hand-curating very large high-quality data sets. Mhm. Which is the Scale AI core business, right? Yeah.

And so another paper, let me put it on screen here really fast. This is the self-adapting models out of, um, out of MIT. What they're showing is that these, you know, they they have a great analogy here. So, it's like a a student that goes to to school, reads all the, um, uh, the textbooks, read, you know, all the lectures and then writes their notes, writes down the notes, kind of compress all that information to the notes, and then, um, kind of using those notes, uh, studies off of those notes. And that's very, very effective. And so, what's interesting is they kind of are doing that, but the models are also able to, um, through supervised fine-tuning do these self-edits and change. Yeah, change their weights. Um, you know, so basically changing kind of like the the weights like the how it it how the brain of the model works so to speak to be fine-tuned to do some specific task, doing it by itself, which is seems really interesting because, you know, one of the things that we kind of talk about is that maybe AI agents, you know, the autonomous AI agents aren't quite just around the corner yet because these things tend to fall apart over long horizon tasks; they don't have that long-term consistency and coherence. Um, and part of that is I think one of the reasons is the models are static, right? So they're, you know, they're sort of they're trained and then it's like, you know, going to work and then on day one you have a certain you have a certain brain, right? And it doesn't change, you know, on day 100. You don't update your information knowledge, uh, other than the stuff that you're able to write down. So this is a little bit more fluid it seems like in in a sense that it can in real time update its knowledge base and its weights and its abilities. Um, so yeah, can we talk about this paper because it seems kind of like a like a big deal, right?

Yeah, this paper is even scarier. I haven't done my whole work on it. Joe or Wes, you have. Go for it. I'm happy to talk about this paper.

Yeah. So this paper is using a very similar idea as the previous one we were discussing. It's like you're asking the model to give you, uh, some feedback or some training suggestions, and then as you said, Wes, uh, then you're actually going and doing fine-tuning on the model weights themselves, and then if you're thinking about RL training as needing a verifier or a reward signal, your reward is after training the model, like modifying its weights, you run some tests on, right, that are specific to the thing you trained it on, and you see how well it does, and then that is the reward that comes back to the RL side. So you're sort of training two things at once. You're training the weights of the model on some example, some new data that you want it to understand, and you're training the original model on how well it can suggest training parameters and examples, which is pretty crazy. It's like you're training a model to understand how to train another model. And also you and I talked about this before, Wes, but that's a long cycle. I mean, even though they're using a very small model, I think it's a billion parameters, which is relatively small, uh, it still takes them something like 60 seconds to to do the, uh, LoRA-based training update and then run through a couple of tests, which is a long time to wait for an RL reward signal.

Yeah, I I missed that part where but it sounds like the amount of compute is fairly large to do that even on small models.

Yeah. Mhm. Yeah. And you can imagine if it was a large model, you might be waiting minutes to get an update even for training one new data item. So this is a long cycle. The other thing that's scary about this method is it sort of suggests why only do a LoRA update. And notice they throw away the update after they're done, right? They really just want the reward signal. But after you're done doing many such samples, you could collect all of those and use it as a training set if you wanted to update another model. And furthermore, you could start asking the first model not just for, uh, hyperparameters for doing a small training run, but like maybe changes to the

Model itself, like, can you suggest, you know, how many layers should I have, or how wide should the layers be? Or maybe we're going to modify the attention mechanism. God knows what changes you want to make to the transformer stack itself, which is something normally that an RL researcher would do, or an ML researcher, but you could ask the original model those kind of questions if you wanted to. And now you'd have an even longer period before you got the feedback.

Yeah. Yeah. And one of the researchers—um, I noticed on Twitter, and I think you had it in your notes too—um, did say that the final idea, yeah, is like the teacher and the student model. So you separate them out. All that data is used to train a model that's better and better at suggesting stuff. Um, and then, um, yeah, I mean, I'm seeing more and more stuff coming out that's similar to Alpha Evolve, right? So the idea is you have these large language models as kind of like the pilots, and then you have like a various scaffolding, scaffolding around it. The model throws out a bunch of—it's—I mean, the trick seems to be how well we can gauge, how well we can evaluate the outputs. If we're able to test, or the Darwin goal machine is kind of the same thing, right? So it's able to improve its own um abilities to code. So it's like, if you're able to evaluate the outputs and then do some sort of that evolutionary tree search, like, "Oh, this cluster or this lineage of ideas seems to be working really well; let's think through that lineage," and like that seems, seems, it seems to be working incredibly well if we're able to evaluate the um the final output. I just saw another paper where they they did the same thing with um uh with—it's playing Settlers of Catan.

Oh yeah, and it's it's it it got pretty good, and it's testing a whole bunch of different stuff. If it's doing research online, finding strategies, and it's like you can probably apply this to—I mean, a lot of things, certainly, and we're probably going to see more and more—just not copy and paste exactly, but they're going to take that idea and just apply it to so many things. Um, I think Dr. Jim Fan and the Nvidia team were one of the first people that I saw doing it with Nvidia's Voyager and their Eureka and stuff like that. I mean, when I first saw like, "Wait, this thing is improving—it's like improving its ability to train these robots in a simulation just by sampling a bunch of answers from GPT-4" was at the time.

Yep. And I remember at the end of the paper they're saying how like, as the task difficulty gets better, not only does it get better than humans on some of them, but also there's like this divergence between the ideas that humans come up with and what the model comes up with. So it's almost like these novel approaches that we can't necessarily think through, uh, you know, come up with. I was like, "Man, you know, if nothing stops this progress, this seems like there's going to be a lot of very interesting applications."

Um, anyways, so I think the the Anthropic uh papers talk about how close they are to being able to automate the work of an ML researcher, and I know OpenAI has also mentioned that same metric, because what you're hinting at is if they can demonstrate a system that can handle those kind of tasks—the tasks done by an average ML researcher on an average day—then you would get a sort of takeoff where their automated systems would augment the effort of their own team members, and you would—unless like you said—unless you top out somewhere, unless there's some diminishing return, you would just get this ramp, uh, and the systems would just keep improving. You know, everyone's left the building.

Yeah, it's absolutely—yeah, the the OpenAI, they have their ML paper bench, I think it is, um, that literally like, can they replicate—so if you give them some PhD paper about a machine learning experiment, can they replicate the codebase and run that experiment, confirm it? And it's like it's not quite there yet, but man, it it seems, you know, getting better and better. So, at some point, it's going to cross that line, I feel like. So, um, well, this this uh self-adapting language models paper is a definite step in that direction.

Mhm. And and their suggestions at the end are sort of hinting like if they do another year or two of work, they'll be another big step in that direction. Yeah. How how far away do you think we are from actually uh being able to handle average tasks for an ML researcher in their team? So that's that's a very—I mean, I I have no idea. Obviously, I I deferred to you guys about these things, but I mean, the point is I think that for everybody listening, everybody uh um on the live chat right now, so because you guys call out BS when you hear it. So, if somebody has some crazy idea about where it's going to go, what it's going to take, you know, you guys are like, "No, here's why," and you you have very good explanations for for for it. But I mean, as as as we're hearing now, um, the idea of a some sort of a takeoff once AI research is automated is not some crazy scientific pipe dream. It's could, you know what I mean? These papers are beginning to suggest that we might be getting closer to it. I think I read uh uh Sam Altman's um recent paper, the gradual singularity or gentle singularity.

Yeah, gentle. Gentle, the nice the nice one. Yeah. But the friendly singularity here, back better than ever, rebooted. Start the branding early. You got to like—Yeah, it's fine. It's uh—Yeah. No, but one phrase that kind of uh uh jumped out at me, he he called it we're at the larval stages of, you know, recursively. That was so unfortunate. We use Laravel. I'm like, "Oh god." I mean, I used to play Starcraft, so I'm thinking of Zerg in my mind. Are you saying we're going to get infested now? Like, what's going on? But I think it's kind of a a neat way of looking at it because it's like, "No, we're not there yet," but we're seeing a lot of these things that seem like if we keep pulling at that thread, it's going to unravel. We're going to, you know, with Alpha Evolve, all this other stuff, it it it's all kind of—we're approaching that recursive self-improving AI. And yeah, maybe we'll hit some limit, right? We don't know. Maybe, well, even even assuming that you do uh see diminishing returns, even assuming that it's not an exponential, it's an S-curve, that's that's okay. An S-curve can still take you a long way. And when you do hit the limit at the top, it can reveal other potential improvements. Like we thought we were going to top out with just training of these large models, right? Everyone's freaked out that GPT-5 isn't, you know, imminent and so much better. And same thing with Claude 4 and so on, not to mention Meta uh for—But you know, we had test time compute, inference time, right? Came in and it was another scaling sort of environment, and we're seeing huge improvements there still, and I assume that will top out as well—that at some point if you give the model uh you know an hour or two to work on something, you won't get any big improvement over giving it a half an hour, say, right? I don't know what the exact limit is; that's okay. It's another uh independent way of scaling the capabilities of the model, and this sort of thing that we're discussing now might be a third or I don't know what number we're on now, but another way to get interesting improvements, and even if it tops out, it'll probably reveal other ways, right? And there's, you know, law of diminishing returns, and bottlenecks will always appear in some different way, and we'll figure out ways to improve upon it.

I think what a lot of people are doing too is trying to sell this um panacea of, "Oh, we're going to hit this point where there's going to be no limitations ever for humanity, and we just put on our 3D glasses, eat popcorn, and just let the AGI take off," and for this whole um ML researcher, automated ML researcher—I use these models for me—an indicator would be, okay, are these AI labs now slowing down hiring, or are they maybe paying less for certain roles? Because I mean, you're paying a million, two million, $3 million a year for some of these researchers. And that could be money that could be going somewhere else in your company. Um, and you know, it might not necessarily work out that, you know, there could be that Dr. Mike, who we super appreciate, he could say also, well, no, Jordan, what could happen is now you want to hire more AI engineers and use these models to make them even more efficient or something. And I think the contrarian would say, "I don't know if we want to 50x our codebase and make things more complicated here. Maybe throwing more bodies at the situation is not the answer." So interesting to see how that all that will shake out. Um, I think we're hitting our 115, 120 limit, and we got to move on to part two.

Yes. So um we're going to continue this uh conversation on the SVIC podcast. So uh it's it's still going to be the three of us. We're just continuing it. We're going to do a little bit more Q&A. So everybody hang on for just a second. We're going to redirect this thing and we're going to continue because next we have—maybe you know what, we spent so much time talking about this stuff, which was very interesting, but I think maybe even some of the more interesting stuff is coming up because what we wanted to talk about is DeepSeek China. It sounds like the new version of DeepSeek might be resembling the Gemini model a lot more because before it resembled OpenAI; now it seemed like it's um resembling Gemini a lot more. We got a paper where somebody—not confirms, but basically breaks down their approach to train DeepSeek, the GRPO I believe versus the the PO, and and um some interesting findings there. So we definitely want to cover that. Uh, we have that paper. So stay tuned. Don't do—Don't go anywhere.