To admit that you are worried about software taking your job as a software builder is to effectively say that you are a bad software engineer. But if it's as broad based as you guys suggest that 200,000 people in this town are not going to be here anymore. Why aren't 200,000 people talking about this? Because it would be to admit that you're the 200,000 that's getting fired, not the 200,000 that's going to get paid more.
Why hello friends, welcome to More or Less Indian Wells edition. I'm here in Palm Springs. You are, we're not there. At the Information House, Indian Wells, the hottest place in this tournament. But the news is so hot, I had to break from the festivities to be with you all. How's everyone doing? I like sporty Jess. Oh my God. There's a lot going on this week. Thank you for being here, Jess. It's nice of you to take a break from your fourth hour of tennis in the last two days. To tie on the podcast. Guys, you know, there are some good players. Do you get paid for this? This is, have you figured? I do. This would be a great time to thank IBM.
She does get paid for this. This is Jess's favorite job she's ever had.
I'll call next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next next Golden Lager. Okay, great. We all have our sponsors. And I'm wrapping Racket, which is also all over Indian Wells.
Latest edition just hit newsstands. Do you remember when you worked in the advertising business? Yeah. I remember, and I'm still traumatized by it. I love, I'm here with our corporate subscribers as well. So I'm in all the businesses at the moment. And I've got to say there is a tech connection because as they call him, Uncle Larry, as in Larry Allison, owns this whole operation and is doing a pretty good job sitting over at Nobu right now. So I actually don't know if he's there at this moment, but... Wow, did he just get a free plug in the information? Well, while his son acquires Warner Brothers. But you know what?
This is my favorite asset that is owned by Larry Allison. I've got to tell you. Fire Sam Altman is trending on X again. Is this like October of 2021? Yes. What happened? Let's get to the news. Let's get to the news. Okay. Why is AI so dramatic? Because they made it dramatic. You get what you ask for. Minutes ago, my friends, the information broke the latest volley in this Anthropic OpenAI Department of Defense war. And the latest memo broken by my team was that on Friday, Dario, the CEO of Anthropic, wrote a 1600-word memo in Slack, basically defending why they had walked away from the defense deal
and blaming the mendacious Sam Altman for undercutting their reputation in public. And while writing that email or Slack message, he also said that Anthropic had failed to curry favor with the Trump administration by not acting with dictator-like praise of the Trump administration. So that is literally the breaking news. We published first some excerpts from the memo and then the whole memo. And I'm really interested in what you all think about this. Obviously, I think there will very likely be blowback from the government that behind the scenes has been continuing to do some back-channeling
and negotiations with Anthropic about getting their technology back in production in some way. I think this might make it harder. But what's also interesting, and people should read the whole memo, it's really saying Sam is trying to undercut us, but now we're trying to undercut him. And it illustrates the whole real culture and employee wars nature of this. So to your point, Sam, about buyer Sam Altman trending again, we are back in that part of the cycle where it's like winning the hearts and minds and employees and culture of these companies. And I think it's interesting because we've been here at so many points in the AI cycle, and that's my takeaway.
But what do you guys think? Well, beyond all of that, Jess, I think you have to mention that the consumers are also super upset, right? So like chat GPT on installs, we're at 295%. There's this quit GPT movement, fire Sam, like you mentioned. And then meanwhile, anthropics revenue has doubled. It's like 19 billion ARR. So the bigger question is... Wait, over what time period? Over... What is that, Jess? Like a week? Not over the last two days. Does it have anything to do with this? It has nothing to do with this. No, they say he was speaking at Morgan Stanley. Well, people are pointing towards it. Basically, there's this question out there in the field, which is like,
is it cool now to say no to the government because that can earn you so much marketing cred with the people? Yeah. No, this is classic face versus heel marketing. It's like, anthropic is like taking this marketing moment all the way to the bank. They're the principled, you know, actor that is doing the right thing, standing up against the man, face, face, face. Like, and OpenAI is the evil heel, you know, da-da-da-da-da. Like, that's like, this is the ultimate marketing moment for them. They're playing it like a fiddle. Remember, before all of this, there are the Super Bowl ads, too.
I mean, so it's not just that they're seizing this moment, but this has been their playbook to attack. Yeah, this is clearly their strategy. This is clearly their marketing strategy. Or are they just like trying desperately to stay in the news cycle when other stuff is going on? Like, I got to say, I'm like right now on Twitter. They're not happy about this memo, so. I'm liking, I'm enjoying Senator Sanders, Bernie Sanders. Will AI become part of the humans if snow is humanity? I went to Silicon Valley to ask some leading AI experts first. Bernie Sanders is running around San Francisco asking people AI questions? This is wonderful. What great Bernie Sanders.
But isn't Anthropic the AI, or isn't Anthropic the Bernie Sanders of AI? No, because it's good. Well, but they would be more politically aligned. I mean, this was Dario's in the Slack message, which I thought was really interesting, because we see all these like highly polished treatises on the future of humanity from Dario. Right? Like he is the philosopher CEO, always publishing. So it was really interesting to see like his raw internal Slack messages, which actually weren't that polished. And that's also interesting to me, because what CEO sends a 1600 word Slack message in 2026 and doesn't expect it to be out there?
But I think, Britt, to your point, I don't know what to make of all these uninstalls, climbing the app chart thing. Right? I assume that just kind of sorts itself out. Right? Like, remember, delete Uber? Did anyone delete Uber and then like keep Uber deleted? Same stuff. I mean, I don't know. Every time this happened at Facebook in the early days, it was the best thing that ever happened to us. Remember when Life360 got banned by the teens on TikTok and everyone gave it a one star rating to just lower its app store rating? And it did. It suffered for like a month. And then it became the best thing that ever happened to it. And teens actually really like love Life360.
It turns nuts back to like 4.8 now. So, yeah, I could see this reversing. I don't think this is long term. I mean, is it the bigger question that the product, like if the products are comparable and there's no differentiation, you know, I mean, that's really the bigger, bigger question. And this is happening against the backdrop of the walk up to all these IPOs hitting the investment banking conference circuit, all touting their ARR.
You know, one thing I will say from the conversations I've had here, there's a lot of skepticism about these AI companies, enterprise ARR numbers, because people are pointing out they're selling these huge seat deals because companies don't yet know how to buy this stuff. But eventually all these companies are going to say, I'm not going to buy, you know, however many thousand dollars seat for all of my employees. I'm going to want usage based pricing or I'm going to want some kind of outcome thing. And people are just very skeptical that the model that they're selling, AI companies are selling at the enterprise today will be the model in a couple of years.
Just like with everything with AI, you're just speed running all of history all over again, right? Like experimental budgets to real budgets to like buy whatever. It's like everything just speed running, you know? So, yeah, like there's no way it's not the idea. Like the whole thing with SaaS, right, was everyone was like, oh, this once you acquire a customer, they're infinitely valuable because they will never cancel and subscribe forever to this thing. And so like they could like spend all this money. Like, but that turns out to be bullshit, right? And like the same thing will do with this. It's like there's just general intelligence. It's low margin.
It's I mean, it's great. I'm like literally building stuff as we talk. I'm like into it, but I think it's a terrible business. Well, that that is our second topic on the pod. So dear listeners of More or Less, Sam Lesson has been busy this week replacing us all with bots. And I'll let Sam talk about what he did. But you can also go to our website, moreorless.com and click on chat in the top right corner. More or Less pod. More or less pod.com. What is moreorless.com? Is it a porn site or something? I don't know. But Sam, why don't you tell everyone what you've been up to? I mean, so like every six months, I just spend a lot of time vibe coding.
And I will tell you that this time it's different because it's like truly a thing you can do instead of internet shopping slash looking at Instagram on conference calls. Like it's no longer like almost... Do we need to short Instagram shopping revenue? Like are we worried about that at the moment? No, but I will tell you that I've bought fewer things this week on Instagram because of vibe coding, but not for the reason people think, which is just like it's so fun to build stuff, right? Like I was joking the other day, it's like, it really is at this point, which it wasn't even six months ago, that this is just like roll blocks for adults. So, you know, in the...
I've built like six apps this week. Like literally, I'm not even exaggerating. But like I did first build a quick version. I basically took all of our transcripts. I ripped through them, had to build personality, psychographic profiles of all of us, like the patterns of like how we chat, etc. Then I had it start pulling... I built a little thing, mallchat.ai, which you can get to from our website. But it basically... The first little applet was pull the news once an hour. Jessica, as the moderator, picks the topic that matters the most. And then we have like an open-ended debate, the four of us, like in our words, in our language, you know, on the real-time news.
Then I was like, well, I can just make the entire podcast. So I had another thing I built out using some stuff I worked on before, which is like top news of the week, pick the topics, do a full podcast. And then by the way, turn it into voices. So I like we're gonna have with 11 Labs, all of our voices, you can literally just auto-generate pretty credible podcasts. And then I built a chat version where like you can just come in if you're signed in and ask a question to all of us and we'll respond to four characters. Then I built a flattery bot that just uses our characters and like does an entire like... It says just say nice things about me.
And so I built an entire podcast of just an hour of nice content about me. We need the flatter bot to use our Twitter to flatter everyone in Silicon Valley. Or the troll bot. Maybe they should troll everyone instead. Like if we have the face or heel option. No, no, no. We're gonna flatter everyone on more at more or less. Flatter bot honestly is a thing. I'm gonna keep working on it. It's great. It makes you feel awesome. And it comes from... I had a real life flatter experience, not on purpose, but we were doing some like... I had a real life experience where we were doing some stuff for an announcement that's coming up. And so we were talking to a bunch of entrepreneurs
and saying, hey, just do us a favor and give us kind of like some nice things you remember about us and our interactions and found... And so you basically... I had two days of people being like... And LPs saying nice things on camera with me. I'm like, this feels great. I'm like, wait a minute. I can just automate this. I can have flattery all the time, right? From characters that are in my life. And the nice part is with your real voices, it sounds great. So the flattery by is the least. So Dave and Britt, as hearing yourself and your voices engaged in this podcast, what was your reaction? Like, was it one of these, oh, no way? Or were you like, could you...
Have we crossed some chasms? My voice is definitely the best, I think, of the four of ours, I will have to say. I'm not sure if that's a good thing or a bad thing. My voice is the most easily like replicable for bad purposes on the internet and our authentic future or for my advantage in this scenario. But I'll take the under on that and say like, I'm into it. Like I thought we... And it also, by the way, got the tone of what we all talk about, right? Like our personalities. So I appreciated what you did this week, Sam. Thank you. I'm not even done. I've got more stuff I'm working on as we talk. But it's, yeah. Can we take like giant archives of files
and start publishing like parallel podcasts using our voice? Oh, that was my question too. Like if I wanted to do a Taylor Swift podcast... It's so easy. Could I take like an archive of her life, have AI sum it up into 13 episodes? I've got open search backends. I've got the whole thing firing. By the way, incredibly easy to do at this point. I can take millions of records. I actually, you know what I did? I took all of our transcripts, ripped them apart, had them in a Postgres database by character so we can remix and redo anything we want. Like you give me any text file, it's like literally three key strokes and I can give you anything you want.
This is like a really great gift you can give someone. You know, you used to like give someone a cool AI... When AI was new and you could like make them a cool AI photo of themselves because only a few people knew how to do that yet. Now, Sam could be like, Britt, you're like a really good friend of mine. I've actually created your AI digital identity. I've got your voice right, your personality right, all these things. You're welcome. And like give it to me in a package that I can... I'm happy to give that to you, Britt, and whoever and all of you. And like, honestly, it's just like, it's so fun. It is the most fun with computers you can have while on a conference call.
I gotta say, my reaction to the first kind of transcript, first topic was like, this is totally a conversation we should, we could have. But it's very one note. Apparently, I only care about governance issues. So whatever the topic is, I'm gonna point out now. And I would say like maybe one out of five times, I think there's an interesting follow-up governance question, but it really has me pigeonholed into that track, I gotta say. Well, yeah, but that's in all fairness, that's just because like I did this really quickly, right? Like I can, I'll get it more sophisticated for you. I mean, I also care about business models. I also care about the bigger picture.
Sure, these are other Jessica-isms that the AI could have picked up. I can tell you, the Jessica, here's the Jessica, like what the bot, again, this is like compressed. It has other elements to it, but your psychographic is analytical, journalist-operator hybrid, likely ENTJ, three forward profile. Like it has your- You added that. No, I didn't. I actually honestly didn't. I asked it to do a psychographic, but I didn't hand code any of this. Sam, what's your upshot on all this for our listeners? Yeah, I wanna comment. Do you want us to release an AI-only pod? But is it like, this is what's happening? Like, where is this going? Like, what's the takeaway?
No, my belief is that this is so fun. And like the entire dead internet is gonna look like this. And it's super fun. And I love playing with it. And there's no value in any of it because it's way too easy to make. There's no value in listening to this podcast. My thing is like, it's, it kind of just like all this content, code is just content. All this stuff is just content. It's fun. Some of it might be really good. To Jess's point, some of it's kind of useful. But this idea that like any of this stuff is valuable is, I don't see it, right? It's like, it's incredibly powerful and tends towards zero because like everything will just work this way. And like,
I'm like maybe a few months ahead of people just because like, I'm just technical enough that there's a few hurdles that, like for instance, I did a bunch of work to set up Jessica bots for her company, which are sick. Like they're so good. And they're like- Like what is it Jessica bot do? Ask her editor when the article's coming out? It's a lot of, a lot of analytics and then a lot of hopefully, ultimately marketing suggestions, very tailored marketing suggestions around individual articles. It's so good. It's so good. There's a stuff we were doing six months ago you had to like manually patch and like now it just like understands how to like, and so, but it's,
it's just too, like no one's gonna build a company doing any of this stuff because it's too easy to build, right? Like there's no software defensibility to any of it, right? Like, so you just have to be really pure on business model and like why it is that you're gonna have a compounding advantage versus like you did something that was quote unquote hard that a customer asked for and that's somehow valuable. I mean, this is, my number one rant, I've had several meetings this week where I feel like people are still stuck in the old model of building companies where they're like, I'm like, what are you working on? They're like, we did this. I'm like, why? It's like, well,
we talked to the customers and the customers said they wanted X so we built it for them and now we'll sell it to them. I'm like, no, you're not gonna sell it to them. They might want X. That's very nice. That's not customer discovery. Like that customer discovery, YC style, like blah, blah, blah, blah is just so wrong for the moment in terms of how to build a business because just because like bill in sales has some CRM shim problem, it doesn't matter. You can't build it and give it to him and sell it to him because anyone can build it for him in a second and give it to him and he'll be able to build it for himself in four days. You know, like it's, you know,
it's like that's game over. Now, I will say from an infrastructure perspective, my God, is my digital ocean bill gone up. Not just the anthropic. What is going on in the pool? I don't even, I luckily, I am spending a legitimate amount on like infrastructure personally now that I wasn't before. So like those numbers will go up, but ultimately like I just use digital ocean because it's straightforward and easy for the things I need. Like it's not, like there's no margin in it long term because it's very easy to switch. Okay. Well, dear friends, I'm going to turn over the moderator mic to Britt and go do my other day job in, in chatting with the great corporate subscribers
of the information. Get us some sponsors for the pod while you're at it, Jess. You know, I could, people love the pod. So I wish you luck in the remainder of this conversation and I will see you all soon. Thanks, Jess. See you. Bye. All right, boys, you're mine now. Get excited. I wanted to, I wanted to raise another theme that we've been seeing Apple this week announced they've got new MacBook Pros and a MacBook Air and an M5 chip and they are citing the RAM shortage, but they, you know, they're saying like, this is now a local AI machine. And so further cementing the idea that Apple is potentially the winner in all of this. What do we think? Did this help their case
this week? Dave, over to you. I mean, this was, everybody's been waiting. I think what's actually, most people are disappointed about is that these chips aren't in the Mac Studio and the Mac Mini yet. You know, it's great that they're in the laptops. It's like they chose the wrong devices for what everyone needs them for. I don't know. It's super clear that Apple is focusing on inference at the edge. I mean, we've been saying this, I've been saying this for, you know, I don't even know how many podcast episodes now, but, you know, this is their unique advantage. And so it'll be interesting. I mean, one of the things I heard from somebody I was doing
this week in startups podcast earlier this week and they were talking about running, you know, a lot of their inference tasks locally on a Mac Studio. I haven't started doing it yet. I want to try it. Apple's got a serious advantage here. I mean, this goes back to my, like, again, I hate Apple, but I respect this move where it's like, everyone's like, we're going to spend $100 billion on infrastructure. And Apple's like, you know, who's going to spend $100 billion on infrastructure are fucking customers. We don't have to pay shit on it. Right? I'm like, ah, that's pretty good. That's pretty good. Make the customer pay for the computers. They're the only one
executing a business model. They're like the only ones in AI, other than, I guess, Google, that's executing a business model in AI that actually works. Sell the computers. What a crazy concept. I don't know. Don't go off balance sheet for like a bajillion infinity dollars in debt. Do the thing you're great at. Why would we buy computers? We should sell the computers. Genius. All right. We also, since we all gathered here only a week ago, we also watched as Jack Dorsey did a 40% cut of Block slash Square, as well as companies like Microsoft, Salesforce, HubSpot, et cetera, all starting to watch AI agents start to erode their user licensing business models.
And even the HubSpot CEO proposed toll gates to monetize agent access to the data. So it seems like there's continuing to be this existential threat that AI is taking over in an agentic fashion as well as within the workforce itself. And that Block perhaps is the first of many to come in terms of cutting back staff because of the productivity gains. Agree or disagree? I just think it's funny that like if you think about the two companies, Jack, it's like Twitter and Square, right, or whatever it's called. What is it called? Block. And you're like, okay, so like Twitter, Elon did it for him. And cut that from like thousands of people to like 200 people or whatever
ridiculous number it now is. And he's like, ah, I probably should do this one myself. Right? Like, it's a cover story, but it's also, it's a cover story because like- It's definitely a cover story. But at the same time, it's like, it's also true that like I, AI, like why would you hire most people? So- You think it was both? Look, I think it's a cover story in the case of Block. They might as well take the stock boost on like whatever and bigger story than like, if you said, oopsie, we overhired and fired people, you don't get as much credit as you say, we can do it. You know, you get some because you still fire people, but like you get more credit to be like,
there's some agent, future AI story. I will say though that like, it was interesting. I was talking to some public markets investors that this week and they're like, I think this block thing's a bigger deal because it signals, because they were saying like, we talked to really big companies. Like we see no AI impact on our workforce. And he's like, people have been saying, ah, this is going to like change the narrative on that. And I, maybe that's true. But to me, it's like, yeah, like obviously these companies are all going to shrink a ton. And like, I don't know, Dave, you think Silicon Valley, you still believe that Silicon Valley isn't going to fire a ton of people?
Because like, I'm the guy, I mean, we're all 50% cutters and you think everyone's still going to have a software engineering job. Well, you and I said half of the software engineers that are employed right now in the Bay Area will not be employed three years from now. And Dave said, no, you guys are wrong. It would definitely be more. And then this happened. I'm just saying. No, you guys are still wrong. We're still wrong. And you're, and you're playing a semantic game of how to label it, blah, blah, blah. There's no possible way we aren't building more software in the future than less. Like that future is already here. Sam, you built how many apps this week?
Like the future is already here. But with how many people, this is the point. It's like, just because there's more writing doesn't mean you have more scribes. Just because there's more calculating doesn't mean you have more calculators in like the hidden figures sense. I understand. But like, I don't, they may not be working at the Mag 7, but like people are, there are still going to be technical people building software for like Cambrian number of things in the world. Like, it's just like not going to be the case that we're doing less. Am I a software engineer? Yes. Seems like the product managers are the big winners here. I, well, okay,
I guess that's the semantic thing. If you consider me a software engineer. I don't consider Sam a software engineer. That's what, that is the, Sam is technical enough. I mean, he has never been employed as a software engineer. All right. Uh oh, did we get over Dave's line? We pushed him over. it's just like, this is just like a semantic difference in like, how to think about whether or not there are going to be people making software for other people and what you title them. And like, I think that's like what this comes down to. And we're like constant, like you're trying to like argue semantics with me. And like, I'm telling you that the number of people making software
for people in the world is going to go up. Like it's just going to happen. No one's arguing that. Taking photos. No one's arguing that. The number of people who, I think that's the same thing. The number of people, for instance, taking photos in the world has gone up dramatically, but the number of photographers has plummeted. The number of people who type is through the roof. The number of typists has gone to zero. So, I mean, this, this, there's this question, which is like, okay, Hollywood had a massive strike about this. Right. And I mean, it was like biggest strike in the history of Hollywood. Right. Related to, I guess it was name and likeness, writers, et cetera.
Right. All related to AI. The question is, why haven't Silicon Valley software engineers gone the same way? Instead, they're just like driving, they're like building the gun that they're holding to their own head is what you guys are arguing. What would they strike for? What regulation do they want? I'm not talking, I'm not saying strike, but nobody's talking about it. I guess our friend Aditya wrote like a piece in the information today. That's kind of. He did? Bemoaning this. Yeah. He wrote a piece in the information about, you know, that software or writing code is dead. And I tweeted it this morning. It's just, it's like kind of strange to me that everybody
just seems okay with this. Like people are just like, like what, why is that? Two reasons. One, unlike Hollywood, Silicon Valley is not unionized. Right. And like, that makes a big difference in terms of the culture and things like that in terms of, in terms of all that. You've often made this, you've often made the argument that maybe they should be. I agree. Well, well, in former world, they should have been. It's too late now. No, but here's the, here's, here's too late now. Here's the thing about, I think the more important element beyond just the structural unionization part is look, you're a software builder to admit that you are worried about software,
taking your job as a software builder is to effectively say that you are a bad software engineer. Right. Because it basically, the story should be for the best of the best, right, which by the way, the best of the best is going and getting paid even more, right? Like, I am one of the best of the best. Therefore, I am not worried, but you should be worried as not the best of the best. Like, it's kind of an admission that you're not good at your job, right, to some degree. And so I think while a lot of people actually practically do face this and feel a lot of angst about it, in some ways, like, there's this thing with software which is not true of Hollywood
because it's a separate industry where, like, if you are of the chosen few, you should be happy about this, right? And I think that dynamic keeps people from, like, speaking out very much about this because by being worried about using your job to AI, you are almost implicitly identifying yourself as someone who is not, doesn't think they're the best at what they do. That's, I think, the issue, right? So it's like, because, like, the story is actually true, right? But if it's as broad-based as you guys suggest that 200,000 people in this town are not going to be here anymore, like, why isn't anyone, like, why aren't 200,000 people talking about this?
Is it just the social? Because it would be to admit that you're in the 200,000 that's getting fired, not the 200,000 that's going to get paid more. But I think there's some middle ground, right, where, like, it's not that they're not going to be here. I don't think they're going to be employed as software engineers. But that's what you guys have been arguing me, arguing with me about, that they're not going to be here anymore. I don't think they're going to be employed as software engineers. I think they're going to, like, use their engineering brains to develop other skills. And, like, I think there will be new jobs that have different titles that they might be.
And so they might still be here, but just in a different role. I don't think that's the spirit of the question you've been arguing. They might not get paid as much. I'm like, I don't think, like, these software engineers, well, think about it. They've been kings. Like, I worked at Google. You guys were at Facebook. We worked at Apple. They've, like, named the number. Like, they were just, like, running around these places, like kings, anything to get the engineers happy. You know, it was so easy to switch jobs. A lot of them actually had, like, 16-month, 10 years. That's the average of a software engineer in Silicon Valley because they would just
cherry-pick lottery tickets across all the startups over the last 15 years. And now, it's like, uh-oh. You're saying that stopped? That's still going on. For the best of the best, sure. I don't think for all of them. Like, I think we're in a, like, really interesting time where a lot of these people are starting to think, like, uh-oh. What is my role five or 10 years from now? Yeah, I definitely think they're seeing that. I will say, my funny line on this right now, which you guys remember, is, remember, there's, like, a famous old saying, which is actually true, which was kind of, like, Silicon Valley, like, how it worked for a long time, which is everyone saying,
code wins arguments. Like, this was, like, this big thing. My line on this is, like, huh, interesting. If code wins arguments, but now everyone can write code, then who wins the arguments? As opposed to it being, like, the engineers win because they can, they wrote the code. Isn't that why people are so excited? Because they're going to get to win arguments? Because everyone can speak the same language. Yeah. Yeah, it's just the people who used to win all the arguments, like, are, lose leverage. Yeah, and what, what becomes valuable is, like, that he, you know, he or she who can communicate the best idea wins the argument, right? And now everyone can write code. So,
you know, like, it's a more level playing field. Everyone can look at what they might, what the outcome might be. I'm not promoting this. I just think it's a massive shift and very destabilizing. I will say, like, I think there's a land grab right now for the engineers that are smart and entrepreneurially minded to be, like, starting companies and actually filling the gap that they used to have of, like, needing to find, like, a business minded co-founder or whatever to be, like, their agent while they, like, can manage fleets of agents. Like, I think there's, like, something opportunistic in the water right now for those, for a lot of those people. Maybe, maybe,
but can I define it a little differently? But here's what I think is opportunistic because I think what's really amazingly cool is effectively the IC engineers who weren't managers, maybe who didn't communicate well and can just, like, manage. So, like, what the real opportunity is engineering management getting flattened because I can just do so much more myself. I don't need to talk to anyone. The business thing is interesting because I actually think this is, like, one of the big angsty things right now is, like, look, if you were an amazing scribe, right, in, like, the year one or a thousand BC, right, like, that was great but you didn't know anything other than
scribing. If all of a sudden everyone can scribe, like, you have no comparative advantage. Whereas, like, if you know a business really well or whatever, there's still, like, some, maybe, let's talk about the depth of expertise, like, there's some angle or some community you have trust with or you have relationships, like, there's other assets. I think the problem with being a pure scribe or a pure software engineer is, like, you don't, that doesn't give you what you need to be successful at business, right? Like, I do, I do think it can knock out engineering management if you're a great IC engineer but not a great manager but I don't think it, like,
I actually think this is a real problem with a lot of people whose, like, core advantage or, like, if you were a computer in the year 1920 and really good at adding, it's, like, you don't have to do that anymore, right? And that's what you're specialized in. Maybe, but, like, I mean, we have this company I think we mentioned last week called Pulsia that's just, like, on a tear right now as a one-person company that's gotten from, like, zero to two million ARR in two weeks and, like, the agent, the agent is just running the company. It's doing all the, it's in 150 investor meetings in the last week. It's done, like, all of this stuff, like, making strategy decks, like,
managing financials, like, all of that stuff. But I think that's, like, quite interesting if you're a technical person who doesn't know how to, like, run a business but now you can. Yeah, I hear you. I hear you. All right, let's change gears to more exciting topics. Our friend, actually, Josh Wolf at Lux Capital wrote an email to his founders at Lux saying, we are at peak AI. This is also in the information. What a week. We are, the bubble is about to pop. Everyone, try to go, like, try to go public, get funding, like, batten down the hatches because the impending doom of the recession is coming. So, now everyone's obviously debating, was this an overreaction
or is this warranted? I think it's just a way to get, I mean, it's a totally reasonable opinion. It's either right and he'll look smart or it's wrong and no one will remember, so who cares? Do you think that this is going to infiltrate the rest of the valley? I mean, like, Lux is a really well-known firm, right? So, do you think others are going to draft off of this and be like, yeah, like, we should, we should be, like, battening down the hatches here. This is definitely going to be recession. There's a lot of groupthink in venture capital, so I'm curious how the halo effect will take shape. I guess I'm, like, trying to look up what the points that he's trying to make.
So, I guess, all right, he says, bonds and equities are telling different stories. The 10-year yield hit a four-month low while stocks sit at near-highs. When these two diverge this sharply, fixed income tends to be the more reliable signal, sometimes a recession predictor, so that's point one. He says, physical infrastructure is seeing political interventions, so 300 data center moratorium bills have been filed across 30 states. So, grid interconnection timelines are, you know, stretching out, so there's a bunch of risk that growth plans and compute get stalled, I guess, by political moratoriums. That's point two. Point three is tariff structures have shifted.
A new balance of payments framework puts structural floor on import duties. I guess that's what Ryan Peterson and all those guys were talking about last week. You know, many of you import hardware, materials, and those costs are fluxing all over the place. So, that's number three. And number four is what we just talked about, which is the markets rewarded blocks 50% workforce reduction with a 20% gain. And so, he's suggesting that those four things suggest something's really off and a recession might be accompanying. What do you think, Sam? Yeah, something's clearly really off. Both things can be true at the same time, right? Like, we're in like a major
moment of disruption in like a bunch of different ways. And we've got a hot war. I'm surprised you didn't mention that. But that's good for the economy. Yeah. Most people don't understand that, I think. Also, like, it's good to get rid of terrible, terrible regimes. So, congrats, good job, America. By the way, my favorite thing about war, I have to say, our military's very good. It's like, I saw this like CENTCOM update and they were just like, here's what we're, like, very like, matter of fact. What's CENTCOM, Sam? Central Command. And like, they just like, the guy running it was just like, oh, I don't know and I should probably know who he is. Just keep like a very,
like, to the point, clear, articulate, this is the situation, this is what's happening, this is what's happening, da da da da. And you're like, competence is just so attractive. I'm like, thank you for being competent. It's like, especially juxtaposed to like, all the blowhardness that we're used to in the internet and a bunch of other places. Like, the military is just like, competent. And like, I know there's all sorts of problems and I talk with smart people about procurement and wait, like, I get it all, but like, there's something very comforting and fundamental about just like, competence and execution. And like, I think, you know, go America, go Israel.
Seems very competent, what's going on right now, at least from the military side. Also, it's kind of cool that we blew up a ship with a torpedo since first time since World War II. Yeah. And clearly, all the AI defense companies are benefiting from this as well. I saw Saronic just raised at a $7.5 billion valuation. They make warships, Andrel, et cetera. So, do we think this is, I haven't actually paid too much attention. Like, is this going to be a quick one or do you think this is going to last for a while and does this therefore net impact, like delay, whatever recession? Come on, Sam. Not our lane. Give me your war times. Definitely not our lane. What did CENTCOM say?
I just enjoyed the competence. I mean, I look, I'm just happy that we got rid of a terrible dictator who is awful and causing a lot of terror. What comes next? I understand people who are like, understand, like, it's always complicated and I understand people who are upset about blah, blah, blah. But at the end of the day, like, you have to be willing to act, right? Like, if you're not willing to act, like, you have nothing. And like, I applaud our willingness to act. You know what it is? It's kind of like, you know, I was talking to some entrepreneurs all this day, but like, I think this is kind of a Silicon Valley narrative that's been built up
and people keep repeating, which is something like, the future belongs to people who just do shit or like take initiative, right? I think that is 100% true. It's like, when you're looking for people to hire, you're looking for people to work with, it's like, the people who are going to just take initiative and like, use the tools at their disposal and like, do stuff are the people who are successful and get things done. And like, it's crazy to say this, but I think this is like, is true in like, things like global politics as well. It's like, I applaud the US just doing things, right? As much as anything else. What's our net net on recession in 26? Look,
I've been saying for like, I could argue we're already there. I look for two, we're not mathematically there, right? But for two years, I've been saying like, look, there's this like triangle trade between DC, Silicon Valley, and New York. Whereas, huge deficit, huge debt problem, huge debt problem, GDP has to grow massively fast in order to deal with it. It's like the only way out because we can't cut costs to make it happen realistically. And so the triangle trade becomes Washington being like, GDP must grow to like, Silicon Valley being like, I have a magic answer called this AI thing to make GDP number go up to New York and Riyadh and wherever else being like, cool,
we'll fund it and we'll get in the game, which is really exciting. And so this is all, there's like, there's this huge narrative collusion around like, our only way out is up through AI and maybe space, space and all this type of stuff. And like, look, I'm partially a beneficiary of this. I can't complain too much, but like, that's just a very clear thing. And the question is, what breaks that narrative cycle, right? Or what, what breaks it? Because that's what's propelling everything right now, right? It could be, oh shit, it turns out that like, GDP headline number sort of goes up because we build a bunch of data centers or spend a lot of money, but actually it's not
growing the economy in a more holistic, healthy way. It could be, oh shit, this actually causes a bunch of jobs to go away and that's a huge problem for consumer spending and blah, blah, blah. Like there's so many ways it breaks, but the real question is like, you have this narrative that's been very, very successful and like, there's really cool stuff coming out of it to a point, right? But what's going to break that narrative? And I don't know, I don't know when that happens or why it happens. There's a thousand ways it could break. It's always been a bit of a narrative story, but it could keep partying, you know? Like, I don't know. No one knows. If we knew,
we'd trade it. Well, isn't there some, I mean, don't we need, we need from a geopolitical stance, AI to keep being the narrative so that we can throw a ton of resources into everything we're building so China isn't like continuing to overtake us with everything that they're doing. I mean, some of the models that they're creating are really good. Well, this is the thing, actually Dave, I wanted to ask you about this because like the interest, there is, the China story is quite interesting. I was thinking of this, I saw this thing pointing out that like the place that like open cause, crushing it is China. Yeah. Right. And there's like, and it kind of couples in my mind
with the open models. It's like, ah, interesting. Like I'm kind of using Claude code. Like I think Claude is better for what I want. Again, I do run open Claude and like have it running, but I'm not using it anywhere near as much as I use Claude in like my Claude instances. I pay a lot for these models. I'm like willing to pay more for the tip of the spear. But then there's all this like down market open stuff, right? Which is like largely from China, largely because China's blocked out of the really fancy, like the best stuff. So they've had to like be scrappy and figure it out. And like, what does that, what does that stack? That stack is like a bunch
of the open source models. It's open claw. It's all this other stuff effectively. Like what's your talk track on that? Like is open source AI a Chinese phenomenon for the, for the foreseeable future? It's an interesting question. Um, I mean, I've been hearing that the Kimi model is really incredible and the team that built it's really amazing. I mean, I think there's this interesting question I tweeted about earlier this week, which is that if you can give a lower quality model a longer energy budget, does it even matter if you're operating on the better model? Because I'm hearing a lot of people that just like give, you know, a longer reasoning budget
to a lower quality model and it comes out with the same answer anyway. And so, so there's that question, right? Like does, do the chip, you know, the export controls that we're putting in place for the best chips even matter, um, in the long run? And by the way, do they even matter for being the best in the world at this, at like any given task, right? It might just be that if you have like infinite cheap energy, you can just like let it reason for a really long time with a lower quality model. And so, what you see coming out of China is a lot of these like open source, lower quality models, right? And them trying to like spread it and get a lot of people working on it
to make them better and better and better. And like, you're seeing a lot of people do that, right? And you're seeing a lot of people with open claws running Kimmy because they don't want to spend the money on four, six, right? And they're gonna wake up their claw in the middle of the night five times to do the reasoning. When it's cheap. Yeah, exactly. And so like, I think that's like a, that's like a real thing. Yeah. And so for me, it's like, look, I even with my claw, not with claw, but with my claw. Like I just use open router now, which is great by the way. Like open router is wonderful for like all the things, right? But what do you use it for? Do you route amongst
different models or, I mean, all open router does is proxy calls to the different big models. Well, it can also balance between them or like optimize between them, et cetera. Yeah. I just, I just, honestly, I think it's a better interface. Like it's less brittle, I find in a lot of ways. You let them handle all the weird connections to different things, right? Like now, I do think like long-term, it's one of those places where like, look, back in the day, when I started my first company with drop.io, like we had this whole vision about routing storage to the cheapest storage providers. That turned out to like not matter or be a thing because it was like, nah,
it doesn't that much. I forgot, I forgot you guys were doing that. It doesn't, like it was the right mentality, which is like, this is just a commodity. I don't care what backend it is. We're just going to like route to the cheapest storage option. It didn't end up mattering that much. It was more intellectual than practical because like the difference in cost wasn't that much and it wasn't like fast enough. This is the version that actually does make sense, right? We were like, okay, I don't care where the intelligence is on earth. Just route to the cheapest intelligence based on what I actually need. Like don't, don't give me a fucking Einstein when I'm asking a problem
that a two-year-old can do. And like when I need Einstein, like make Einstein work, you know, like that's like a good mentality and a expensive enough thing to do. But it's just interesting because I just wonder if we like, the closed world is clearly the American world ironically right now. And like this stuff is good. It's like really good, right? It's better. But to your point, especially if we end up with like economic jitters or all of a sudden these labs can't pay people infinitely and open source moves faster and it's already like fine. It's just going to be an interesting thing if that turns out to be a Chinese phenomenon, right? If like, if the world becomes
US is Apple and Claude Code and expensive shit and like the Android of AI is open source Chinese shit. Not just Chinese, but probably rest of world. But it's mostly Chinese. Partially in AI, I think you're doing open source if you're behind, right? Like it's a distribution strategy. And I think Sam, if you went back to like maybe the beginning episodes of our pod, we were talking about Meta running a distribution strategy because they were behind, right? And so they started putting out Lama and all of these open source models to try to just get the distribution and get everybody in the ecosystem working on them because they didn't need to monetize the models.
They just needed to, like they can take whatever the community makes better and then put it into their advertising system and make money rain from the sky, right? And so you could argue that China is just running the same strategy here broadly. I buy that. It's just an interesting moment. To train what on? Where's the money coming from? No, no, no. Not to train anything. It's just that, you know, they have, you know, less chips in volume than we do, right? And they are therefore behind. And so they need to get like as many eyes and hands working on their various technologies in the AI realm. And so they get out there faster if they use open source because people share it.
A hundred percent. But it also just becomes really interesting when, if you think about like, okay, like right now you're talking about the, it's a loop to the economy question. Economy hot, infinite capital for these like head labs with questionable economics long-term, right? But like it's fine. And like they can run faster. If all of a sudden they can't run as fast because they don't have the same infinite access to capital, like how does that play out, right? Like there's all of a sudden, like you know what I mean? Like all of a sudden the edge. But does it become, I guess like, but my question is like, does infinite access to low cost energy matter more?
Yeah, it probably does. Right. And this is the reason why, yes, but that's kind of it. They can put up a nuclear plant and like something like, what is it? Like nine months? It takes us six years. Yes. And so like, they're going to have access to lower cost energy. Like there's like under no metric, whether it's number of solar panels, number of nuclear power plants, like they are going to have the cheaper energy based on how things are going right now. Yeah. And that might, that might be all that matters. in the end. That makes a lot of sense. I think that probably is right. So we need to focus on nuclear plants. Well, people are. I just think it's like,
we're going to have structural problem with that. Like this is just the thing about why I like find this intelligence thing and this business such a weird one because you're like, I don't understand. Like this is basically all going to back to energy at some level. There's not going to be that very much margin in it. And like low cost intelligence production is going to be the key. So Sam and Dave, I guess, as we start to wrap soon, where are you most excited about investing right now as things have changed and continue to change, I guess, knowing that Sam, you, you, you, you haven't probably changed your tuna. How much have it changed? I feel like in February,
February was like a wild month in the state of Silicon Valley and AI. And I think a lot of people are. I think we're in the second wave. AGI has been the name of the game in so many conversations I've been part of right now. It's just like, oh, this is actually going to happen and happen soon. And like, what does this mean on a faster time horizon and blah, blah, blah, blah, blah. I just think we're in the second wave. The second wave, which is what? Yeah, I think we're in the second wave and personal AI or agents, however you want to talk about it, is here. And there's going to be a lot of interesting things to be built. And there's a lot of infrastructure
that hasn't happened. There's a lot of agent-based business models that people are going to explore. Memory is still like this entire area that has, is not figured out at all. And so I think there's just, yeah, it's like exciting. I think that's, that's why everybody's so excited right now. I think that the, it's like, agents might actually be the website of this, like of this era. And like the browser hasn't been built yet. And, you know, we're like at the very beginning. All the infrastructure just needs to be created. Yeah. Yeah. It may be that like, protocols and so many other things. There's like no standards for, you know, sol files and config files. And, you know,
there's no protocols for how personal agents or clause should communicate with each other. Like there's just like a lot of stuff to be figured out. But you don't need standards. It doesn't matter anymore. Sam talking with a protein bar in his mouth. No one needs standards. Oh, wouldn't Jess, if she were here, wanted us, want us to end this by talking about David. No, no, no. I'm gonna, Dave. You blew my, you blew it, Dave. All right. I'm gonna go there. I'm gonna go there as we wrap. Sam, I'm gonna hand the mic to you to do the big announcement. What's the big announcement? What did David's protein pivot into as its new product? Because I think the people
are gonna be excited about this. And if you don't know, David's, David's is the protein bar. I've eaten two of them while we've been chatting, which is like 40 grams of protein. That's a pretty good podcast episode. We can get 40 grams of protein. Ice cream. David's. That's the best news ever. Ice cream. All right, ladies and gentlemen, 30 grams of protein, in a pint, 30 grams of protein, 260 calories, two grams of sugar. Are we more or less? All chemicals. On David's ice cream. I mean, I'm clearly gonna eat a lot of it. Like so much cancer. Whether it makes you sicker or not, who knows, like, is what I would say. Like, it's like, I'm gonna eat it and it might kill me.
Okay. I think we should have a more or less meetup ice cream, ice cream social, as they might, as the kids might call it, and do a trial of all this ice cream. So, more protein for everyone. I'm surprised there's no creatine or anything else in this fiber, but I'm sure that's coming next. It's hard to put creatine in the thing. It destroys it. You gotta put that in a powder and just drink it in the morning, like an adult. Ah, okay. I've also heard about creatine overload. All right. We, dear listeners, are going to head back to our wild lives as being venture capitalists. Because I've been on this call, I've had an API issue with a SendGrid key and I've been in a loop
for like literally like 10 minutes with my bot. I've just fixed it. We're so sorry. We're so sorry. The problem, this call was too engaging. It distracted me from fucking around. That's the goal here, everyone. All right. Well, if you enjoyed this episode, please like, rate, and subscribe and we will see you back here next week for another episode of More or Less. Bye, guys. See you later. If you enjoyed this show, please leave us a virtual high five by rating it and reviewing it on Apple Podcast, 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.