OpenClaw vs Meta vs OpenAI: The Personal Agent Wars Heat Up

Transcript

People seem to want the cheapest or the personality they prefer. The OpenClaw community, people actually choose to route to Claude, particularly Opus 4.5, because they like the personality better than, say, Codex. Maybe, but Dave, that's like imagining two accountants and one slightly prettier than the other. It's, you'll use the prettier one a little bit, but you're just going to use the better one or the cheaper one, I think the better one. Well, you're a rich Brit, so you'll use the better one. More or less Why, hello, friends. Welcome to More or Less. How are you, Morans? Hey, Jess. Doing well. How was lunch? We were having lunch in the same restaurant, as usual. We were talking capital. What else do we do? Talking about capital. Well, this is an exciting week to be doing the pod because there has been a lot of technology news. Stuff going on. Very exciting. We're going to dive into it. Some of the people on this pod are even involved in said technology news. Some of the people on this pod broke the technology news., it has been... It's always busy. I don't know. But it's an exciting time. And of course, we're talking about the OpenClaw, OpenAI news. But there's a lot of other things. Information reported today that Meta's smartwatch plans are back on. So we can do a wonderful poll about who is going to buy a Meta smartwatch, considering I still don't have any smartwatches. I'll probably be on the negative side of that. That's really confusing to me. Yeah. Yeah. No, it's against brand. But, I try. Do you have smart anything? I am trying to be less connected to my devices. So I also have watches that I like to wear. I do not have an Aura ring yet. Although I'm thinking about getting one because it tracks stress levels now. It does. But also, I want to make sure we have not circled back with three VCs and a journalist on this pod. We have not circled back recently to the fundraising market, to the state of startup investing. There have been a number of major rounds that are in the process of closing or have closed. Andrel, the information has new reports out about obviously Anthropics. So I want to get to that too, because,, the people are here for the VC. Speaking of VC, those watchers will notice we're down to VC. I just have to explain to the people that it's snowing in Jackson Hole. And then they will understand why Sam Lassen is not on this late, late afternoon podcast taping. But he likely will come on. He'll have a grand entrance. So just stay tuned. He'll jump in. Don't worry. He'll probably be wearing a beacon. He's here. He made it. Welcome, Sam. All guys, start rolling. I have things to say. That could not have been more perfect. No, we already. What do you mean we've been rolling? I just explained why we were down to VC and then you popped on. So I have some bad news, mostly for my wife, but maybe for Dave and Brittany. Well, I may have just Dan corned myself. What? You broke your collarbone? I'm going to do the podcast and we'll see how it feels. No, we're not doing the podcast. Guys, let's pause. It's fine. It's fine. It's fine. I do not want to do a podcast when you may have broken your collarbone. It's fine because it'll be funny and we'll get to check in next week. It's fine. It's fine. We told you we'd make a grand entrance, everyone. So here's the deal. Sam, does this mean you're not going to Alaska? Don't you want to go to urgent care and take care of it so you can go to Alaska? An hour won't make a difference. If it's a broken rib or whatever, I'm going to have to do whatever. Here's the upshot. I did the classic thing, which is I was so proud of myself. I was on the top of four pines with 15 minutes before the podcast. And I had a great run. It is an awesome shape, like blower pow all the way down. And then I was scooting across the lower hobacks in those huge fucking icy moguls. And I just like... Which we skied five times today. Yeah, it's fine. I was just trying to go a little extra fast. And I went over the handlebars into my pole, into my beacon, into my rib. And I was, ow, that hurts. And then I woke up. I was, oh, my arm hurts. And so then I just skied home. And I was, this isn't too bad. But then I would take my ski boot off and it definitely hurts a lot. So we'll see. We'll see. Sam, this is not taking care of your body. These injuries, like the swelling in the next half hour could go crazy. And we're back. We're back. From urgent care. Do we want to give the prognosis, Jess? We'll give the prognosis. Sam did not break his shoulder, his rib, or his collarbone. So he's in significant pain, but it's not broken. So anyway, dear listeners, thank you for that. Now we are ready. Let's say it may or may not show up, but it will have nothing to do with his perfectly intact bones., okay, let's start with open claw. So the information reported this week that the founder of open claw was going over to open AI. The announcement was quickly followed by open AI announcing that the news was true. And interestingly, that new foundation for this red hot open source project that we've been talking about nonstop on this pod called open claw. This foundation would be established. And who would be on this board of said foundation? Well, none other than Dave Morin. So we are honored to be in the presence of open claw foundation board greatness. Dave, what can you tell us about this exciting project? Don't laugh. The claw is the law. The claw is the law. Thanks for the good intro, Jess. Yeah,, the really good news for the open claw community is that the open source project will be managed and looked after by the open claw foundation. We're in the process right now of getting all of the proper structuring set up. And yeah, it's just super good news., I think it's great for all of the developers and builders and hackers out there., this is the as near as we can tell the fastest growing open source project in Internet history. There's a lot of people building on it. And so maintaining the open source project itself and the ethos of that is really important. And,, I'm really grateful that Peter has trusted me with helping him figure out this part of the equation. I think,, being part of it from the beginning of January, late December,, we've been talking about a lot on the pod. It's just been a really fascinating experience. And I feel, I guess, lucky to reapply some old skills that I have in developer platforms and ecosystems from the 2000s to a new era. So it's been fun so far. You don't feel old, do you, Dave? No, we don't use that word. Never., and again, I know there's a lot of inside baseball, Dave, that you don't want to talk about, which makes sense. But obviously, Peter Steinberger, the founder, going to OpenAI to report, to report, to build more personal agents. It just seems we're, on one hand, this is totally validating for the personal agent idea and the concept. But we also now have this race between OpenAI, going hard at it. I reported with my colleagues and the information that Meta was making a big play for Peter to join as well. Meta matched this news within 24 hours with releasing a new, or talking about a new Manus agent coming from an acquisition they did., Britt,, is this full on,, what are we to make of this? And, which agent is going to win if,, we're at this point where everyone just seems to be moving very quickly to have very similar products? Well, I think it's interesting that Manus was able to launch that so quickly, which means, clearly they had been probably working on it throughout the OpenClaw negotiations. I don't know. Just wondering. Yeah. Oh, definitely. Well, probably since they acquired the company. Yeah. But I think the difference is,, Manus is in the cloud, And OpenClaw is literally on hardware on your own computer. And it's a thing that you own and Meta doesn't own. And so I think,, the people that are going to leverage Manus AI, fine. Sure, it's an agent. And I think you're going to start seeing a lot of these types of agents out there and new platforms where you can build your own agent. The question is, do you want to own that as an individual on your own hardware so that you for sure have full control over it? Or do you want to offload that into the cloud somewhere that another company owns? And I think that was the power and is the power of OpenClaw and why people are loving it so much. And it doesn't make sense for anyone else, like one of the big,, the big tech companies to have a competitive version of this because it's they're not and inherently has to be open source, has to be distributed. And so that's the big OpenClaw advantage. Yeah,, perhaps only Apple has the technology and the ability to deploy personal AI to,, the edge, I guess, onto people's personal devices, because that's their business. But Britt's right., I think the entire ethos of OpenClaw is to give people a way to own their own data, to own their own personal AI experience, and to pull it completely out of the cloud. And,, that's not going to be something that everybody wants or needs. I think there'll probably be billions of people using cloud-based versions of this. But,, for the people to do, which it seems like there's quite a lot of people out there that are interested in owning their own data, then the open source model really matters. Yeah. And I think the big question, too, will be, who can deliver a really great consumer experience? Because privacy will win over some users, but probably most, the history of the internet has been won over by convenience. And it's just early. I still,, Jessica, check-in, coding developer, whatever we're in, now that I'm probably eight days into my AI coding journey. I'm just using co-work and OpenClaw, like constantly. Interesting. What are you building? And sometimes I forget. Well, Dave, I think I'm doing, I'm building an app every day. It's our kids' ski break, so I'm on vacation, still working. But I'm building an app every day. I have not built my app yet because I had to go to urgent care when I was going to build my app. So I might have to build my app this evening. But it's everything from, I did some really cool analysis of years of information subscriber sentiment data that has to do with showing correlation to how public stocks move. And newsflash, turns out there is a correlation between the sentiment data and what public stocks, tech stocks do over the next three months. So that's not an app per se. That's analysis. But I built a social media app that my team is now using that I uploaded online, and I'm looking for more ideas. But anyway, my biggest problem with these things is I forget what they're doing because I go off and do something else, and I come back, and I'm, wow, that PDF is done. Oh, yeah, I asked it to do that. So we're going to move on. We've talked a lot about bots, but we had to circle back to note this development. We've reached a new interesting phase. I am also just so tickled would be the wrong word, but the meta-open AI rivalry is rearing its head in all sorts of ways. We also saw today that Charles Porch, a senior, what would you call it? I should say his disclosure as a friend because sometimes people get mad when I don't say that. But Charles, who ran Instagram, some of the most important relationships, including a lot of their celebrity relationships, is going to open AI to do a similar role there. So I think we have now two very, very recent data points of how this rivalry is rearing its head. And obviously, there's been a lot of poaching back and forth between the two companies. Jess, do you have any speculation as to why Charles would make the move from meta to open AI? I don't, but I think it's what's interesting to me about it is we saw the poaching of, it's the poaching is enter is like gone beyond AI researchers, too. So I think it's like a very interesting book. A new wave of poaching. A new wave of poaching. And there's been plenty of that already. But it's interesting to me., these companies aren't. I think you can always tell a lot about a company from who it sees as its greatest rival. I remember years ago, maybe a decade, it was probably a decade ago. And I don't even know if Sundar was the CEO of Google yet. When did Sundar become the CEO? No, it was shortly after I left Google, maybe like 2013., Charles was at meta for 15 years,, that's a really good run, And you could argue that open AI is entering its 2010 meta era, No, it is. I guess the point I was just saying, zooming out from one move, as I remember when Sundar wrote a memo about Google's biggest problem shortly after becoming CEO, or maybe shortly before. It was all about Amazon. And at the time, you would not have found a single article that had like really pit,, everyone thought Facebook with social was Google's biggest problem or something else, And here there was this memo being, no, Amazon, which was getting into the ads business and absolutely has become one of Google's business biggest rivals and now in cloud with Google Cloud. And so I think if you want to be looking six months ahead to what companies are doing, what employees they're poaching and who they see themselves as most competitive with, which is often not where the public narrative is, I think is really interesting. So now the public narrative is probably caught up around open AI and meta,, it wasn't obvious. And so I think it's like the exclamation point on that rivalry, if you will. And we'll see where open AI goes with ads., we have, they're now in beta or whatever, you can see, have you guys seen an open AI ad? I have not. I haven't seen an open AI ad yet. I paid for it. So I'm not going to see it. I have to have two accounts, I guess., it takes years to build an advertising business., just ask anyone. It takes like five years. Yeah. So it's going to be a long time. Or longer., that's actually a generous way to look at it., it takes like a decade to build like a serious ad business. Well, even something that you could point to and say, this is a revenue stream today, and it could become the revenue stream that offsets some of this CapEx or whatever, But that's obviously a next front. And hardware too,, we reported today on the information that Meta has had on again, off again, smartwatch plans. I said earlier in this episode,, who's going to buy it? Are you going to buy it? Well, and we know that you're not going to buy it. Well, I have, I'm such a loyalist with my Apple watch, by the way. Shocker. First of all, all my fashion-y female friends shame me. Dave doesn't know this, and all the men that listen don't understand this. But, women often get shamed in the fashion circles for like wearing smartwatches. Because they're just not that cute., there are definitely cuter watches out there. But actually, all my fashion, well, you just have cute bands. You make it cute. Okay, for normal every day. But, I like to wear it 24-7. What?, I'm tracking all my data. No. You can't wear that thing to a black tie event. No. Okay, maybe black tie is where I cut it off. But anyway, I can dress it up. Anyway, so this is just a problem for women. But, and I'm, if Meta hasn't yet designed great looking eyewear for women. Men do not talk about or care about this. With their Meta glasses. Are they really going to design a cute watch for women? No., so maybe in the future, once they actually design something relevant for me., maybe that should be their strategy. Just to go after women. Well, I've had conversations with the team working on the glasses about like more feminine form factors and styles. I think they're working. They want to just have a broader range. The electronics, though, are the size they are. Just put some diamonds on those electronics. Just put some diamonds on those electronics. They could be lab grown. It's fine. Oh, my God. Yeah. I'm good with it. We'll see. I, I, and then, and then,, we have the open AI Johnny Ive hardware plans that who the hell knows what they are. No, they've, they have been leaked. They've been leaked like nine times. Was this AI? What's the latest leak? I don't know. Was this AI? You tell me, reporter. This is reminding me of, I want, I don't know what year it was, but do you guys remember the Internet of Things era? Of course. When everybody was talking about the Internet of Things. Huge, huge speech. What year was that? I don't know. It was like early 2010s, Yeah. And I remember somebody went to CES that year and they walked around CES and they were, this is not the Internet of Things. This is the Internet of putting iPads on things. Because everyone was like putting an iPad on refrigerators and an iPad on washing machines and an iPad. It was like a screen on everything. And that's what this all feels like to me. It's, nobody has any original ideas. So everybody's, let's just put more electronics in people's pockets on their wrists, on their face, and we'll see what happens., we'll see what works. Well, we need to brand this moment. If that was the Internet of Things, is this the Internet?, I just feel we need a catchphrase for this agentic world we're about to live in. This could be the moment, Jess, to define the moniker. I assume, I guess I have to get to what the point of this watch is. And I assume it's to get some ancillary power or features to the glasses., because already one model, you have some band. You have the neuro band, I think this can do your fitness and everything else. Yeah. But I think they want to control... They have a product where there's a band that allows,, small finger movements to... Yeah. So... Same with Apple,, going in that direction with the Vision Pro and the whole finger gesture situation. Yeah, I think you're right. I think in the future, everyone is betting that we will have stuff on our eyes, ears, and wrists, hands, likely, maybe fingers with the rings. And we will not necessarily need our phones anymore because we will ourselves become computers. Is that the direction you think we're going in?, I hope not., I'm hoping... In that world? What I want is... I do want more ambient recording. I think I've talked about this. Do you? Yeah, I do., and not just as a reporter, but as a human., I think... I see the value of,, having conversations. Where I really see it in a business context, having meetings recorded, having all of that is super useful for connecting dots and AI and all that stuff. But you as a reporter, don't you feel like the need to disclose everything to anyone you're talking to? Because you're literally having off-the-record conversations all the time. But I'm also talking about meeting., if I'm having a meeting about an upcoming project internally, So it doesn't even have to be... But I bet if I had off-the-record conversations and said, hey, this is off-the-record. Same terms. You've known me for 20 years. It's always been off-the-record. But by the way, I'm really interested in what you have to say. And I want to make sure I understand it and get the most out of it. So I'm going to record it for myself., maybe 50% of people say no, but 50% of people say yes, Because... Yeah. And they might... I would give them the same courtesy, If I were sharing my thoughts. So all of this is going to be streaming back to some agent that you have. Somewhere. That's like your memory agent. I'm like full... Agents everywhere. I know. But this will be like..., you're going to have your memory..., all the things Jess ever said and people she interacted with. And, that just needs to live somewhere,, right now... Yeah. For me, it all lives in granola. And then I export all the......indie files. And I,, port them over. But we're going to have a workflow now where that can change. And then you'll have your CRM. That's,, extracted from that, Because it'll know all the people you talk to the most. It'll probably have, like... You can give it,, personality scores., the people you actually like. So do you think, in effect,, you're going to have a friend ranking list of,, who are actually your closest friends? You guys... More or less thread is,, pinned is one of my top texts, Because it's... Ooh, we're pinned. Yeah. You guys are pinned. I didn't pin you. Apple decided to pin you. Apple did? And so... Oh. Yeah. I didn't pin you. Yes, you did. Maybe I accidentally pinned you. I don't know., this... I'm not the... But anyway, I think all of that... I don't know. We don't... We're going to veer away from the AI rabbit hole at this moment because I want to talk to good, hard, cold, venture startup financing for a moment. Okay. All right. But anyway, the device war is heating up, too. And I just hope we see things that are actually useful and we get out of these,, Kickstarter projects, ? I think it'll have,, this is the moment where it's expanding, Yeah. Everyone's trying something and then there will be adoption. Oh, you wanted to coin it. That's a good exercise. The inner... Is it the internet? Well, I want to coin the internet of things, the internet of wearables, but also the age..., what's happening with the agents? Because we now have agents for everything. It's like internet of things was like when the internet just went into everything. And right now the agents are going into everything, ? I think the concept of recording and memory might become core to this next..., if you look at... If you consider health tracking as the precursor to this wave, which I think it is, and it's the thing that's driving adoption, a lot of these things. I'd just like to point out that this was one of my 2026 predictions. Oh, good. Check. We're only a month and a half in. Right. We can go on vacation for the rest of the year because we got it right. Okay, guys. So startup financings, as you probably know, as venture capitalists, they're like coming hot and furious still. We have not... We're still seeing huge valuations. Latest... Well, we have Anthropic around... What was it? 380 billion? I should look this up. It was something around that number. 300 billion. And,, OpenAI's round is coming together. There's been many people reporting on... You can't call these startups, Jess. No, that's true. Okay. What's the OpenAI round coming together at? I don't know what we've reported or not. So I have to... We have to read the information. A big number. Public information. I think it's... More than the last one. It's not going to be a trillion. Let's just say that. It's not going to be a trillion. It's like two Anthropics,. I think it's... That's my guess. 2.1 Anthropics or something like that. Yeah. I don't know. That's just what my guess is. Yeah. That's probably the right range. Andrel. I know these aren't startups. Okay. But Andrel in TalkStory is at 60. But there are many, many... Well, what are you guys seeing with earlier stage? Are we still in the get a billion dollar Series A valuation mode in the valley? Or where are we at? I think you're seeing the tale of two cities. We've talked about it a lot on the podcast. But narrative capitalism is driving a lot of behavior. Sam would be so proud, Dave., it's just true., there's a lot of name brand founders. A lot of what you just mentioned, Jess, is whatever you want to call it. The private Meg 7 is the new way I've been thinking about it. I love that. I love that way of thinking about it. The private Meg 7 and any founder that can command a serious narrative has the ability to raise at,, a billion plus valuations for very big, crazy visions. And that world seems to be hot. And then on the other hand, there's still a lot of seed stage stuff going on. Seed deals happening every week in AI. And then your typical... There's a lot of incubator traffic out there. A lot of,, open claw projects. The word on the street is something like a third of Y Combinator is open claw already. And they're winter class. I don't know how far through they are. I guess they're... I think it's the first investor meets or one month from now.,, the market is pretty normative, I think, right now in terms of what's going on. I think a lot of people are trying to figure out at the seed stage how to think about what's happening just because things are moving so fast that what you invested in last year......might already be in trouble. And how do you think about what to invest in this year? I think that a lot of... That conversation is happening a lot. Well, and I would double down on say that's not just with the startups themselves., we were at a lunch today or yesterday. I forget. Then there's a CEO there has a 600-person company and is, this is so exciting and awesome. And I'm terrified because I know I need to start pivoting my company to be more agentic and AI-enabled and yet to move 600 people with the types of bureaucracy that we have. And this is only 600. It's not like multiple thousands or tens of thousands like some of the companies out there. But 600 is a big company. It is. But it takes time. That takes months, likely. Not weeks. Definitely not days. And the problem is the landscape is shifting in weeks right now, not months. And so you make a plan as a CEO and you're, oh my god, three months from now, the whole plan was wrong. And that's how VCs are feeling too, In the investing cycle. That person specifically was referencing something we've talked about the last couple episodes, which is this agentic engineering process. Dark factories? Yeah, dark factories where people are no longer reviewing code. The engineering process is moving extremely faster. And if you've got a 600-person organization, shifting to this is non-trivial. So I think this thing is playing out across the stack. So here's a question for you guys. Someone approximated there about... I'm going to get this wrong, but not really wrong. Maybe it's wrong. 500,000 software engineers in the Bay Area? Am I off by an order of magnitude? Dave, the number in the world, isn't it? 4 million in the world? So 1-8 of them in the Bay Area feels right. There's around 3 million iOS engineers. There's around 4 million Android engineers. There might be some overlap. Okay, here we go. But I don't know. I think of it as like there's like around 10 million in the world. So there are estimates that in the San Francisco Bay Area, there are about 400,000. So 4% of global engineers are in the Bay Area. What percentage of those do you think will be employed as software engineers in three years? I think that the labor market drama is vastly overblown. I think that it's likely... So give me a number. Give me a number. I think all. I think that it will be... Okay. I think that everyone's going to continue to be building a lot of software. I think there's going to be more software being built. I actually think that the demand for software..., Jess, as you were talking about earlier and you were telling me yesterday, now you're building things. but I'm not a software engineer. And if you were Microsoft and you employed 30,000 of these, are you really going to keep employing 30,000? I don't know about that, but I don't know that that means that these people aren't going to be managing software projects somewhere. No, no, but are they going to be paid to do software engineering, Is their job... Yes. Okay. So you think... I disagree. I'll take the under on that. I think more people overall will be building software. Yeah. But the number of actual software engineers will need to go down because one great engineer can do the work of many engineers and their job will likely be to manage and oversee some of these software engineering agents at the highest ranking of the company. So there will definitely still be a software engineering job title, but we will not have as many engineers. I just don't agree. And I think people ranging from the HR person to the marketing person to the Charles Porch... I don't even know. I just don't agree....partnership person is going to be enabled to build their own software too. And there'll probably have to be some checks and balances that go on. I get to vote and then we can all do that. I am totally on Brit's side. I think, Dave, there's no way that companies that have software engineers..., these companies rush to hire as many as possible at a range of skill sets. They have really excellent ones that are involved with key parts of their apps and they have really mediocre ones. And these are also companies that are spending way more money than Wall Street wants them to be on infrastructure and they're way cut costs. So I think in three years, you'll be down to... But I guess,, how are you framing this question?, this is really important because... It's a very simple question. If 400,000 people today are professional software engineers... In the Bay Area? In the Bay Area. And they are employed by technology companies to come to work and build software every day. What is a number in three years of people who will be employed by tech companies in the Bay Area to come to work and build software every day? And I would not count... As software engineers. I would not count the journalist who uses AI to copy edit their story as a software engineer,, take the pool today and how many of them have jobs as software engineers?, I'll also tell you... You think, Jess? What's your number? Well, so I will go down... In three years, I'll go down to 300,000. And then I think it probably drops. After three years? I think by then it will drop more precipitously. And I also think... I see 50%. I think the operating variable here is in the Bay Area. Because I think that everyone that understands how to build software is going to have more to do in the future than less. And whether or not they are doing a job in the Bay Area or somewhere in the world, I think that,, everyone that understands software is going to be managing agents building software. They're going to be managing building software for more things, not less things. And I don't know whether they're all going to be in the Bay Area. I like how you threw in our podcast title many times there, Dave. No, no. But I'm using the Bay Area..., Dave, it's healthy to have this debate. And by the way, this is the debate that is playing off across society,, it's just a debate at this point. And your point of view is very reasonable. And it's the optimist, maximist point of view. And I think Britt and I have a different point of view, which doesn't mean that the power of these coding tools won't be unlocked and show up in society in myriad ways. But in the same... I think this is going to be like taxi drivers with the rise of Uber or what happens,, to truck drivers as self-driving cars., I think there will be a disruption. And it's not to say all these people be unemployed. They might go build amazing businesses by themselves,, doing all sorts of things., I don't think we know. But it's fascinating. I would love to hear in the comments from our listeners and watchers about this. Yeah. You should do an information poll too, Jess. The interesting thing might be that both of our takes might be optimistic. If you take what... Yes. If you take what Anthropic says seriously, which,, Anthropic says no one will be doing software engineering 12 months from now. Exactly. And so if that is true, then Jess,, San Francisco is Detroit. Oh, I don't think so. Well, two things. And then we can move on. My opinions tend to be shaped by talking to my sources. And my sources tend to be the people who make these decisions and are thinking about them deeply. And so I actually think I,, and this isn't one, this is a range of people across a range of companies, and they are predicting this level of decline in the job. So this isn't,, my, that's, I form my opinions by talking to people. Yeah, I get it. I think it might be possible that we don't know. No, we really don't. That's why it's so fun. We have no idea. We have no idea. But I think it's so new., Jess, even the conversation you and I were having yesterday about my building, it's so amazing what people are doing. And people are operating at an entirely new, they're operating in an entirely new level of abstraction and software. And so the idea that we're going to make less of it, we're going to make less technology, less digital experiences. It just, I just don't buy it., I think we're going to make more. But those things are two different things, Dave. I think you have to realize those things are two different things, We have, people are arguably, well, I don't, I don't know if I agree with this. People might argue that people are very informed today. The number of people employed as journalists has been slashed over the last decade, you can, you can have an explosion in something while also having this particular profession very hard hit, And, and I think I'm totally with you. I'm so proud. Like even the stuff that I've been playing with and sharing with my team to see other people on the team, just like pick it up and start building their own things. It's awesome., and, and it's, but that's very different from the question of,, how many engineers is Google going to, Google's going to keep their engineers the longest. I think we should all admit that. I don't think Google's getting rid of that. Google's one of the big winners in all of this. So we've said from the beginning, Google just can't fire an engineer for,, but,. I think it's really interesting question. Like if you are, I think this is a really difficult question. Like if you are a major CEO in Silicon Valley, are you really going to let go of 30% of your software engineers, 50% of your software engineers? There's no way over time, gradually. I think the other thing to figure out though, is, we're pretending that all of these places only have the best engineers who are really suited to what is to come next. And the story of the growth of the Valley over the last 20 years has been a mad dash for hiring and a mad dash for hiring software engineers. And I think there's probably a lot of slack in the system for a lack of better word too. And so I think that's why that's also what's going to make this trend hard to really tease out and why investors are so like obsessed with figuring it out because you're going to have layoffs and layoffs and layoffs., you're seeing this at Amazon, You're having layoffs and layoffs and layoffs that Angie Jassy is saying are just getting rid of bureaucracy and then you're starting to have layoffs. But that's a different thing. No, no. But, but also,, I know this is shocking, but sometimes CEOs manipulate the story a little bit. So you can, when it becomes politically correct, AKA Wall Street and your employees will buy it to admit that you're having waves of layoffs because of AI, then it will go from being, we're stripping out layers of management to, we have found efficiencies with technology to AI has changed. Like the narrative will just change. And we're, we're in that gray area now between, but it's going to take,, one, two, three outspoken CEOs just to say, look, I, when we started this company, however many decades ago, we organized it like this. Now we're going to try and organize it like this and they're going to do it and they're going to do it all in lockstep as soon as the tide turns. So it's like who blinks first is my opinion. And I also think it'll be more about less hiring of these engineers going forward. Like if you were going to hire X percent more engineers every year, you probably would just stop that now and get better use of your current engineers that are in-house because they can do so much more. I think you have the right skillset. I think that's right. But, I, and that's where I don't know. Can I also make an argument on the personalized builder vibe coding front? Like now I have all these friends and husbands that are building all these things. But you have multiple husbands. And I'm building all these things. You have to, you have to clarify. You have one husband. Single husband. Single husband. Yeah. But,, you're building a new app a day. Dave's building a new app a day. I'm building a new app a day. All my, a lot of my friends, my like normally, normie friends. Friends, including like all my mom friends, like my friends from Texas. They haven't even gotten in on this yet. So let's fast forward to when all of our friends who aren't super like early adopters are now vibe coding their own apps. They want us to try all their apps because they think it's so cool and interesting. I have to play Sam. We're going to have app overload in our life. And then we won't be able to keep up with it. Why are you playing Sam? Because we have all worked on software systems. we've managed teams of people, worked in teams of people. No, actually, I do. Yeah, no, no, I do now. Your system is very, very sophisticated. And you've been around the building of software. And so your ability to think about what you can build is vastly beyond people who have zero experience. Right. And I guess this is my point, which is that this is still a very technical domain. And people are going to need to be able to,, work with people that understand how technical this is., this whole open cloth thing really exposed this to me, which is, there, this town is really the tale of two cities. Also on this front, the technical people and the non technical people. And the skills that it takes to make this stuff really work, even though it's really powerful right now, are still very technical. And I think that's not going to change for many years., I disagree. So Dave, what I wanted to clarify, I am lucky to run a great engineering team, but I am not technical. And I don't need to be. I literally told cowork to move the template so that the icon appears on the left, not the and to create two versions, one for headline, one for headline and one for body text. And I just followed all of its instructions and it happened. So I don't. Yeah, I couldn't set up open cloth by myself, but I don't know that we will have to return. I really am so interested in our listeners, viewers to comment on if there are 400,000 people employed as professional software engineers in the Bay Area today, three years from now, what is that number? Share your number. We'll remember it. And we'll give you some cookies if you got it right. Share how many months or years you think before your like normie friends that don't live in the Bay Area are going to send you their app that they just made. Because I think it's, I think it's like 12 months. But a lot of the stuff I make, Britt,, so I, it's not clear to me the workflow is to start sharing it though. I think we want to start making random things. Like all my mom friends are going to start making their like carpool tracking app. They are. Britt, they're not going to make apps. I don't think so. I think they're just going to get the answer for themselves. I think they're going to make apps. That's okay. You don't have to, you don't have to back my back. There's not going to be apps. We have uncovered. I need them more in solo more often. I've uncovered some real sources of professional disagreement between you two. So one more quick hot topic before we wrap. So is Sam okay? Sam's fine. Sam, because he's skiing all day, has to cram all of his work into like two hours at the end of the day. And as dear listeners of this pod will now have heard multiple times, we just had a stint at urgent care that really sent us back. So I imagine he's doing business. I would think this is business. We all have business too. It's fine. But,, okay, I want to just flag something that we haven't talked about a lot in this pod, but maybe we'll, we'll leave it as a little dangler. But the debates about government contracting of tech are heating up again. And like all these tech companies that it's been such a big headline, like will the government U.S. Government ban anthropic or not? Have you guys been following this? No, I haven't actually. A little bit. I saw the headlines. I don't have the details. Well, that's it., on one hand, you have everyone competing for the business. And this has been,, of Department of Defense, U.S. Go, whatever it is. That's been true. Always of tech. I saw somebody called them un-American. they, why is that? It's a great question, Dave. I'm not sure I know exactly the answer to it. I think a lot of it comes down to... Are they trying to control access to it? they don't want the Pentagon using their models in certain ways. Is that what it is? Is that what it's about? Like in the one sub headline I read, it sounded to me they have... Anthropic has decided to,, restrict certain aspects of how they use the product. And so because of that... It's a terms of use dispute. Yeah, it's a terms of use dispute. Well, it's a terms of use dispute. Un-American. Oh boy, here we go. Sam, how are you? I'm fine. I told you I was fine. The marital dispute at the beginning of this pod was really cute, you guys. It's not a dispute, guys. Just it was nice how much you cared. How would we have known if he was fine? By the way, you can't move your arm or breathe, but it's not broken. So we're good. Yeah, we're fine. By the way, it turns out that the more or less medical back channel is a hell of a channel. So thank you, Dave, for saying that. We got you. We got you. The best new group chat. Sam, we're wrapping this episode... Great....by just giving a taste of... Oh, let's ask you quickly. 400,000 Bay Area engineers... Oh yeah, I want to know what Sam thinks. People employed as professional engineers in the Bay Area currently. Three years from now, what's your number? And then I'll tell you what everyone else said. Well, did you guys discuss this Dave question of, if you make things with bots, are you an engineer or just a user? Yeah, that doesn't count. Yes, we spent way too long. You have to be paid by a technology company to do software engineering. So it's 400,000 now, and how many years is it in the future? Three years. Three years. 200,000. Yes! Okay. That's Fred's number. Sam, agree. What is... What's Dave's number? Is it like 100 million? All of them. Yeah, Dave was, more of them. More. I'm a more. That's... You are a more. Okay, going back. You joining the last family? Yeah, sure. Will you back to me? You are welcome. I've been really cynical on this pod so far, Sam. You're going to love it. Somebody has to do it. Someone has to fill the void. Okay, so yes. Anthropic and the Pentagon are fighting over terms of use. Meanwhile, XAI and everyone else are tripping over themselves to get these government contracts. I don't know where this goes. Anyone have thoughts? It's just a bit... We hadn't talked about it, so I wanted to put it on everyone's radar. I guess the question for me... Here's the interesting question. Why is this debate between the Pentagon and Anthropic in the public domain?, how did... Who's..., who's... Try a signal chat. Probably Anthropic. They tend to write essays. No, it can't be Anthropic. It has to be the government, That's how they play it, The... This administration. That's true. Both of these are leaker. Well, Anthropic's not a leaker, yeah. I guess the question for me is,, I don't... These types of things don't end up in the public domain unless someone wants them to be in the public domain., if we're arguing about terms of use... It's,, too in the weeds. what ? For it to be a thing. And so I just... That's my question, is I'd be very curious as to,, who's selling the story effectively to back into,, what's going to happen. Elon Musk is probably selling the story, don't you think? No. Maybe. This sounds to me like the Pentagon., they wanted to use the models in a certain way. And Anthropic said, no, you can't. And so they're, all we're going to go take this to the public. I guess I think for me on a lot... I was thinking... I have a tweet I haven't... I wrote while in the backcountry today in my head. And... Maybe that's why you ended up in urgent care. Wow. And,, it's something along the lines... I haven't framed it yet, exactly. Of,, the whole... If intelligence becomes super cheap and abundant, then intelligence isn't worth anything. So... And,, we already have this going on with human intelligence in theory., I was, there's so much more, quote-unquote, intelligence in the world with these machines.,, supply of intelligence up. And because supply of intelligence is so up, unless somehow demand for intelligence increases faster than supply of intelligence, which I think is what some people argue will happen. I just don't believe it. Then,, actually, intelligence doesn't become that valuable. And it's just like a... It's just a commodity that's cheap. You can get anywhere. In which case, I don't really understand why anyone cares., why Anthropic thinks that they can hold the government ransom on a terms of use because just someone else will just give them the intelligence., I don't know. It's, there was this story a long time ago when OpenAI decided they needed infinity money, where the story was race to AGI. Whoever gets to AGI first is the winner of everything and owns all of the intelligence. And it's, clearly bullshit then. Now, really clearly bullshit., there'll be lots of dealers of intelligence. And then the question is, is, if there is, and if intelligence keeps getting cheaper and the cost of tokens keeps getting cheaper,, I don't understand how anyone can hold each other ransom., effectively over terms of use. Because you just like buy it from someone else. It's, what's the marginal cost per token where it does become commodity is the question. Well, I'm just saying, if you think about it, it's, let's pretend humans produce tokens. You'd say humans produce tokens, but they're expensive. You're, why do I, why do you talk to like Claude and not a human? It's because, because the Claude produces cheaper tokens. There's still tokens., you can go to a psychotherapist and get the same analysis that Claude or GPD or whoever gives you. It's just cheaper. And so, okay,, if tokens keep getting cheaper,, one, sure, humans can't produce them cheaply enough.,, that's why you go from 200,000 developers to 100,000 developers or whatever. And, you're just, yeah, you also can't make any money as a company who sells tokens or intelligence. That is possible. We just don't know. There's a lot we don't know, it seems like. Sam, I also think there's like a, there's a social cost of talking to humans for tokens that you don't have to pay talking to bots., holistically, like the whole cost. The whole cost., oh, I gotta like drive., but the other thing. Okay. So, if AI, if intelligence really does end up being like power, Here's the interesting about power. It's super regulated., you can only use PG&E., there's not an open market for power in most of the country, which is how it supports pricing. If there were an open market for power in the country, power would be unbelievably cheaper. But we have like regulated monopolies effectively in power that make these things financially work and finance PG&E or whatever. But people hate them. And like the justification is, I guess all the infrastructure build out and spend that like is physically bound. I just don't think that applies to the internet or intelligence the same way. So unless you're going to be, there's regulatory capture. And if you live in California, just like PG&E, you also have to use open AI. At which point, it's just a regulatory ARB game effectively. If open power, I just don't understand how you make any money on this. Any of it. That is the question., Dave, you're Mr. Claw. Congrats, by the way. That's pretty cool., I don't care what model it uses at all., it's, I just and I don't believe that this is like a devil's in the defaults where like open AI is, well, it's default open AI. So that works for us. I think people are just being, well, I just want the cheapest shit. And it's all the same. As long as it's good. People seem to want the cheapest or the personality they prefer., that's what one of the more interesting things I'm hearing. The most accurate. The best answers. Well, no, in the open claw community, people actually choose to route to clawed, particularly Opus 4-5, because they like the personality better than, say, Codex. Maybe. But Dave, that's like imagining two accountants and one slightly prettier than the other. It's, you'll use the prettier one a little bit, but, not you're just gonna use the better one or the cheaper one, I think the better one. Well, it's also not... Well, you're a rich Brit, so you'll use the better one. I think anyone would want the one that's correct. I don't know. Maybe. And then we're gonna bring this episode to a close because our dear listeners and viewers have been with us for some time now. But it's also not always clear., it's easy for you guys who play and test and compare them all day what's better. And then you use them enough and you get some instinctive feelings, oh, this one doesn't get things wrong. But all of that is really converging. So I think this concept of being better will only matter in the most narrow of use cases. Maybe. But for health in particular, you definitely don't want to be one person off. And I think a lot of people are gonna be using this. But how would ? You're not gonna run the longitudinal study to compare them. It depends., most of the things you're gonna ask for health are, you can ask a nurse practitioner. Then you don't give a shit. But some things you'll be, oh, I actually care about, the fanciest doctor in the world. I just think the routing becomes like so specific. No, I think we are going to be ingesting a ton of our personal data through wearables and other testing,, etc. And we are going to have a doctor as an agent that is taking care of us. And if that doctor is continuously getting things wrong, even on the margins, like that's gonna become a huge deal. And people will lose trust. And so I think that in healthcare in particular, that matters a lot to be accurate and precise. And so, right now, actually, what you're seeing is Anthropics done a ton of investing and like ton of commercialization and, and it's healthcare infrastructure. Open AI is doing a lot right now., there's a lot of acquisitive stuff happening behind the scenes. Which, by the way, makes sense, because they're realizing that base intelligence isn't valuable, It's a commodity. And so then you have to, climb up the... And then they'll,, this is when Sarah Fryer told me that they're looking at licensing IP for, drug discovery, Because the business model, she's the CFO of OpenAI,, and this is something they're just looking at. It's an acknowledgement that like the raw tokens or the raw intelligence to, in and of itself to everyone is not necessarily the great business, That it's more specialized use cases. This is also the story of OpenClaw, which is that it's a great orchestrator across all of the different models. Many, like one of the big changes in the last week is a lot of people now have developed various skills that can route to the different models based on what the use case is in real time to lower the cost to the lowest amount possible, And I think what you're going to see is the rise of a billion bots, That,, cover every conceivable use case. And Britt,, I think your medical case will be,, somebody will create the,, MedClaw and people will route to the right models. The orchestrator will take care of all the problems that you just talked about. Maybe we should create that one day. Maybe. But,,, it's, I don't know. I think, Sam, I think we're keying in on something really important here, which is that the underlying tokens probably can come from anywhere. And how much margin there is in that business over time probably does look a lot like power, like Sam's talking about. And the real value is going to be in the orchestration. Well, it's going to be in something that isn't abundant, is the point., it's,, and it is interesting because I think the big difference, if I was right to this is, if intelligence looks a lot like the energy world, It resolves back to it in some way, shape or form. But it doesn't have the regulation because they're the same borders. It's just like a very hard to understand how you ever make money in it. Yeah. Well, this will be why we've uncovered and left unresolved the big debates. Who knew? I have to do a little bit of a shameless plug for all you listeners out there who like tennis. I and the information are hosting a very cool event right at the start of the Indian Wells tournament. And we are going to have about 30 VIPs. So we are very limited and we are going to hit and play some tennis with Madison Keys, Grand Slam winner, amazing current player, Matty Keys. And Darren Cahill, best known for currently coaching one Yannick Sinner. And then we're gonna have a party and an awesome dinner and very exclusive seats to the start of the tournament the next day. So if you're interested at all, and you don't know where to find me, you aren't very intrepid, but find me. We have a couple of tickets left. Did you see the recommendations my open claw made on who you should invite based on my CRM? I did, but then I was going to put them in my open claw and actually filter them for who I should actually invite. So I haven't had time to do that yet. Because it was a lot. It was a huge list. You want my open claw too, just to add in on the data? Yeah. This is one of the more fun use cases that I've seen a bunch in the last few days, which is people adding their bots to a group chat with me and having their bot explain things and get into the conversation. And it's pretty interesting, actually. Oh, my God. Well, here we go. Here we go. Oh, my God. The Sam and Jessica bot conversation could go in so many. You guys, let's put all of our bots into the more or less chat and see what happens. That's gonna be really feisty. I made my bot send Dave a Valentine's message, but then I forgot to actually confirm to deliver it. So Dave, I'll send that to you after we finish recording. Classic. A classic,, you misforgot a Valentine's Day thing. All guys. I've got to go. I have to pull a Sam this time. All right. I think we all have to go. Dear audience, thank you for listening to this wild episode, which had many threads from the collarbone of Sam lessons still being intact to the future employment in the software engineering sector to the rise of the Open Claw Foundation. I hope you learned something. And we will see you back here next week for another episode of More or Less. Bye. Bye. See you guys later. If you enjoyed this show, please leave us a virtual high five by rating it and reviewing it on Apple Podcasts, Spotify, YouTube, or wherever you get your podcasts. Find more information about each episode in the show notes and follow us on social media by searching for at more or less at Dave Morin at lesson at Jay lesson. And as for me, I'm at Brit. See you guys next time.