The Wisdom Wall

“a lot of the way that we built software, which was like, we're going to build all these little shims in between, and there's going to be all this open source software and I'm going to use yours and you're going to use mine and it's going to make my life easier. Like a lot of that stuff goes away when these models can…”

“These guys jumping ship proves it actually. Right? Like if, if they were so sure that they were on the path to AGI or super intelligence or whatever we're going to call it this week, they would have stayed with the team and the system that they were already working on. Instead, they chose to take money and move…”

“And, uh, what we do know is that more and more and more regulation of big tech equals less and less and less acquisition of small tech. Which results in the complete and total freezing of the innovation ecosystem of Silicon Valley.”

“These clouds, I mean, I don't think Meta publishes it or Google publishes it, but my guess is that it's around ten million computers in like these clouds globally, if not tens of millions, and no one brain has any chance. Parents have absolutely no chance. And, like, I really think that's what this is about, is that,…”

“The point is if these systems start improving themselves, like less than two years from now, the more cynical take is that if you're one of the 1000 people who have the most coveted of all skill sets, which is in training, pre-training, the sort of raising of models, then you need to get out the door now in order to…”

“this is how you know the cycle is coming, like, at least somewhere in the tail end, because people aren't actually talking about customers. They're only talking about the money that they raise”

“The user experience is light years away from being easy enough for anybody that doesn't understand at a deeply technical level what's going on. And I think that's why you see only single digit millions people using things like cloud code and codex and, you know, the most advanced agent tools because they're just…”

“A harness is the software built on top of a model to make the model useful. The models are not useful by themselves. They need to be given software that helps them think in certain patterns, repeatedly run loops to accomplish tasks”

“A dark software factory is basically you're running a agentic engineering organization where you, the code must not be written by humans. The code must not be reviewed by humans and you write tools to understand what's being built and to keep the thorough put and the, bottlenecks in the speed of your development at the…”

“in the seed market, you have the multi-stage firms, And 2000 seed funds competing for the 75th percentile of all of those sort of just like you said, the mainstream stuff out there in the market. And so that effectively means you have almost perfect pricing and a bunch of synthetic beta in the top quartile of the seed…”

“Truth is social is dead today, right? There, there are, there are no more social networks that anyone uses because everything, you know, followers don't matter. Your friendships don't matter. There's no actual functional social relationships inside of any of these systems today that matter in terms of the signal that…”

“I think that there's this fundamental thing which is hard about AI for Apple, which is that Apple really believes in consistency and delivering a quality experience the same time every time. That's pretty opposed to the nature of these probabilistic computers that we call AI.”

“Whether or not people use it for therapy or for helping them think through things or not, Sam, if you change the personality of the underlying tool or the underlying LLM without transparency, I think that's the most dangerous possible thing that you can do for mental health.”

“You can't, you can't do Apple style launches, um, in the social media era.”

“And so my theory has been that if you take up a bunch of the spots in your mind with people who aren't your real friends, Then we, when you go off into the real world looking for your real friends, you're going to wonder where they are. And so I think that the thing going on where people are saying there's a loneliness…”

“Right now you're seeing almost no consumer AI apps because it's so expensive to use the big models, right? Like if I were going to build a consumer AI app of any kind, call it, let's call it a social app of any kind. And I needed to call that API every single time a consumer interacted with the app. It's prohibitively…”

“The answer is, well, you need to depend on Apple making their small models that are on device available via API, which they haven't yet, right?”

“nation states have enormous resources and spend enormous resources to break encryption, find back doors, like do all kinds of things to make sure that they can listen to your phone if they want to. And it doesn't matter whether you're on Telegram or Signal, nation states will find a way.”

“the hardest thing to do in a network to get engagement is to get the friends to respond to the thing that somebody posts. And so if that's true, then friends in social media will truly die. They, they already are dying, but like this is maybe the death knell.”

“This pattern is the same across all of these technologies, right? Like you've got this really sort of hypey early hacker phase where you, you know, have a bunch of people working on something, a bunch of capital is deployed, none of it works and it all crashes. Then we go through it again and, you know, iterate into…”

“Don't do a half layoff with the hope you'll get to profitability. Do the layoff to profitability once get yourself to default alive and then start making decisions.”

“like the thing that kills most seed companies is actually founder breakups, not the business not working.”

“Our bits, as they flow out of our internet companies, literally hit a brick wall when they get to the Chinese border. And yet Chinese internet companies can, you know, develop enormously huge market caps and monopolies and technologies inside of Chinese borders. They can then take those bits and flow them all over the…”

“we're, we're in the command line phase of this thing, and nobody's actually figured out the graphical user interface for, uh, for these, uh, AI technologies, and we've got a long way to go before we do.”

“in order to run multiple agents at once, your Mac has to manage each thread, right? So if you're running, you know, one, two, three, 10, 100 agents at the same time, whether they're running through Codex, Pi, OpenClaw, or whatever, you have to have memory and compute available locally on your machine in order to run…”

“I think in the longterm, if you can make your developer ecosystem make more money than you do, you are in a much better position than you are taking all of the profit for yourself over the longterm.”

“That the only thing holding you back now is attention and, like, time and creativity. Like, you don't need, like, it's already so powerful that it doesn't really matter.”

“There's no difference between an app that's built on Linux or an app that's built on iOS. Like these are just apps.”

“Partially in AI, I think you're doing open source if you're behind, right? Like it's a distribution strategy.”

“these are actually the most sophisticated and most influential form of social media ever created.”

“It's what all great founders do when they need to, yeah, exactly. Hide layoffs and change the business model.”

“this is actually the Harbinger. Of knowing when the cycle has hit its peak, right? When everybody is sort of talking about how they're sleeping in the office and we have, you know, 24 seven coverage.”

“And my whole thing is like, what makes this any different than any kind of electronic music? It's just a different type of electronic music. Just because it sounds like classic rock, you know, doesn't mean it's, it's still electronic music.”

“these AI Companies and business models that they are like computers. They're not like services. Like I buy chat GPT pro because I want to have the best computer to use for my, you know, everyday life. Just like I buy a new Mac book once every two years or a new phone every one year. I want to have the best compute in…”

“I continue to believe that this is more like the computer business than it is the internet business, where we're all paying for computers. Like, we're paying for access to computers, and in the early personal computing revolution, everybody was, if you're a serious professional, uh, trying to play in the NFL business…”

“Four year vesting was four year vesting because that's how long it took to get a company public in the nineties.”

“For these enterprise SaaS companies, that's all the value that there is, right? At some point, the everything's commodity. And so the entire battleground is sales, right? Like anybody who runs an enterprise, an enterprise SaaS company knows that, right? That like, once you're at a certain scale, that it's all sales.”

“The good news about investing in startups is they're mostly, you know, recession proof. They're mostly off into the future.”

“Like every social network since forever knows that bikini bikinis are the real influencers, right? Like that's like what drives the traffic.”

“you buy a meme coin to become part of a community. And by buying into that community, you make the community more valuable.”

“all fundamentalisms, and we've seen quite a, even, you know, in the last few months, we've seen other forms of fundamentalism make grave miscalculations in terms of the actions that they take in the real world. And fundamentalisms in all of their forms are the fundamentalist, you know, philosophy becomes a cult. Or…”

“people's mistrust, these types of things is largely driven by incentives, right? And the way that people have to write headlines and stories and make videos, which gets the clicks, which gets the money, right? Like it's all this WWE, um, Politic that's going on right now where you have to either be a heel or a face and…”

“It's like in both cases, whether you're in pop culture, Taylor Swift land, or whether you're in politics, they're both just American pop culture of different flavors. And it's all of this stuff is just WWF characters.”

“Actually, most people, I still don't think associate tech talk with being one of the first AI, you know, uh, products at scale that did this first. Now it's just much more direct. It's like, you're talking to the algorithm rather than the algorithm serving up content that leads you in this direction, but they're the…”

“And, uh, what we do know is that more and more and more regulation of big tech equals less and less and less acquisition of small tech.”

“This is actually kind of my point, is that Magnificent Seven are now Actually kind of an IPO. Like the scale of the liquidity in these seven companies is so large that selling to one of these companies is as, you know, the scale of it is almost the same as, I mean, what, Amazon went public for four hundred million or…”

“When an asset class is dot domain. So like, remember when web three was using dot X, Y, Z and dot X, Y, Z domains went through the roof. They were like Like, 10,000 dollars per domain. So like, when that happens, you know you should go short.”

“at some point the scale of the capital stack gets so big that secondary becomes to Sam's point, the only way that people can get the piece that they want. But it also becomes, you know, the gravity for the employee base becomes so strong that you have to start providing it. Otherwise you've got things like, you know,…”

“to play in hardware, you either have to be the most disciplined startup entrepreneur of all time, and probably somebody who's done it multiple times, and is willing to go through it again, right? And to be extremely capital efficient in the world of atoms, which almost no one can do. Or you have to be, you know, one of…”

“The hardware stuff that does work as a startup tends to have an enterprise use case of some kind that enables it to, you know, really have a business model straight out of the gate. And those tend to be, you know, in the case of zip line, it was, uh, healthcare and logistics and delivery in parts of Africa, where,…”

“one of the things that defines this idea of founder mode is skipping the organizational structure. And going directly to the people that do the work. And you know, this is famously how Steve Jobs did it with the top 100 at Apple. It's how Elon, I think, seems to do it at many organizations where the managers are not…”

“Sam, your intuition might be right that like, we've got a big new wave coming and everyone's focused on AI, but actually we all know anybody who's thoughtful knows that AI is just the water. It's like in the water. Now it's going to be in everything.”

“it's very hard to compare what's going on to the internet wave because LLMs are not networks, they're computers, and they're very simple. It's like you put in A bunch of things happen inside a box and output comes out and that's a computer, not a network.”

“You can be, it doesn't matter. You can choose cheers as your KPI or booze and they both work.”

“This is what I, look, I think that you end up with kind of a barbell, right? Where you have a whole lot of AI generated stuff that has an extremely fast metabolism through culture, right? Like this, like lots of stuff coming. It's like, I don't know, it's like sugar food, just like tons of it moving through culture…”

“country is the new rock and roll, and we kind of lost rock and roll. There, there's no real rock You know, genre anymore. And so I think that people go to country because that's where the guitar continues to exist.”

“Yeah, if you're thinking about laying people off, it's already too late.”

“We're in the command line phase of this and like the operating systems and the word processors and the spreadsheets for this stuff have like not even remotely been created yet.”

“I think people are getting extremely sick of Consuming AI slop in their work days. You hear this refrain is really starting to kick up. Do not send me work slop. I literally don't work with humans so that like the human can talk to the AI and then copy and paste what the AI said to me. There's like, those conversations…”

“You no longer have to speak a language that a compiler has to understand. Now you just have to be able to speak a language that the models can understand, and the models go speak compiler, and the compiler speaks to the assembler, the assembler speaks machine language, and like everything's”

“I keep saying this, but I think it's kind of 1997. I always say it's like either 1983 or 1997, depending on which part of the technology canon you want to look at. But we're kind of in that era where websites were just invented. Some of the early languages for programming them are around some of the early servers for…”

“your, your hands can only type so fast and it's just more context per second, right? That you can input into the computer. And I think that's the big change here is that the models want more context to do a better job for you. And so we are shifting our way that we're interfacing with the computer to match the…”

“Because like one of the other things that we're seeing a lot of is that, you know, if Anthropic decides that they need to shift their business model or whatever, whatever intelligence you're using underneath your agent, you still want its soul and its memories and its understanding of you to come with you as these…”

“we're kind of like just moving up the compiler stack again.”

“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.”

“Gen Alpha memes are, like, speedrunning abstraction. They've, like, first we lost, we lost all context into, you know, And now we've, like, lost language, so we've just, like, gone completely into numbers, and when Gen Alpha decides something is memeable slang, it becomes, like, six, seven, or sigma, or skibbity, or…”

“And I actually started to realize like, oh, this is the social network. And this is maybe not rocket science or that insightful to anybody listening, but you know, this is the social network where a lot of these memes are moving.”

“when we were building Facebook in the 2000, we knew we had beaten MySpace when they started selling the homepage to the Hulk movie, and the Hulk started, the Hulk, like, punched through the MySpace homepage.”

“the feeds for a long time were trying to split you into as many cultures as it possibly could, right? Because you can, like, that makes sense from an advertising perspective, right? Like you want people to have as many thin cultures. You want to have a very large portfolio of thin cultures that people are not deep in…”

“I actually think this is your, this is actually a classic technology cycle story around when a new technology cycle starts. There are a bunch of different players in the market, and one of them will kind of choose a product lane that the other one that is the actual leader will just subsume over time.”

“Anytime one of these technology cycles gets really, really, really going, it causes this recursion loop of, You know, everyone in the entire Valley and then sooner or later the entire country. And then like to Brit's point here, like at the white house level, everyone's talking about the thing and it's like talking and…”

“It's very hard to develop Anything that's got anything even remotely looking like a social use case or a, you know, a use case that hits the LLM APIs very, very frequently without having a local model that's way cheaper than the ones that you have to access in the cloud.”

“I think that that's really why people are so excited about vibe coding. It's like, there's this like quieter revolution going on. That's like, people are learning how to express themselves through software and building like little worlds and GPTs that feel like co-founders and like personal apps that do things for them…”

“The modern 20 25 strategy in general is to war of attrition through all the noise, right? Like there's so much noise. You can play a war of attrition almost across any topic. Um, and as long as you go long enough, People are so fatigued by everything else going on that you're going to get away with it.”

“Like they need people interacting with it a lot more for a lot more reasons in order to build a good ad market. It's basically the identical thing we were working on at Facebook.”

“What I'm arguing for is that people do not build only for Instagram. I think you've got a lot of people that are out there building only for Instagram, only for TikTok, only for YouTube, whatever. Just like an investor would never invest in a single stock. You want to be in a portfolio of places. And ideally, most…”

“Well, I think there's this interesting thing where, like, people try to metabolize new releases of things, like, in the first hour after they launch, and something that's 31 songs long, or even You know, new gadgets that take a long time to understand. It's like, you can't Metabolize it and write about it fast enough.”

“On the founder side being removed, that usually comes because the growth rate of the founder is not keeping up with the growth rate of the company. And so they're falling behind in their ability to lead the company and to make strategy changes that need to be made. And that's usually when you see the board take a move…”

“sports media is proving to be the final frontier in the internet age of linear storytelling, of live events.”

“NBC and these, the big platforms had this like static audience that they could tell the story of the athletes to in the lead up to the Olympics. So they could like Like build the story of an athlete for the many months leading up to it. And that's just gone now with streaming.”

“The browser isn't a browser anymore. It's like the walls are, are, you're, you're able to, like, change the walls of the physical space and the floor, and you're able to, like, do things like what we would call a web page in the internet today, but it's like the surface of the world becomes spatial computing.”

“I would argue, Sam, that it has a cult multiplier. Because maybe, you know, web two and social were hype generating technologies where people got really, really excited, and then the technology itself made even more information spread, so people got even more excited. AI has that same factor. The thing about crypto is…”

“And for the most part, the ecosystem uses one contract and there are all of these different forms of risk, right? And everyone goes around pitching all these, uh, you know, limited partners, the same contract and the each limited partner wants different types of risk.”

“It's okay to get to the end of your seed round and turn to your investors and say, the card we flipped over didn't prove what we thought it would. And so we're either going to shut down or we're going to do something completely different. Right. Like that's the thing we've all signed up for in this ecosystem.”

“Well, I think that hardware is, um, it's like entrepreneurship on expert mode, right? Like it's a very bad thing to try to do if you're a first time entrepreneur.”

“I think this is why VCs like Twitter. And why VCs like crypto, and why VCs like podcasting. You know, it's like you see numbers go up every day, and you can get a feedback loop, right?”

“I mean, every social communicator, every, whether it's Yahoo, AOL, Facebook, everything becomes chat in the end.”

“People reaching out for meetings usually have a decision to make, or they are trying to work through an idea, and it is happening in the present. By asking for a meeting and then punching it out into the future, you're actually operating on bad information.”

“Well, I think that's why betting. It's just betting's the easiest way, right? Because betting's where all the tension is.”

“the trends that start, you know, trends that happen 20 years after, they tend to happen 20 years after the, um, the products become, uh, popular because the people that were buying them, and in this case the nineties, are at their peak earning potential.”

“people really desire this like tribal connection, uh, to other people and to have a shared experience in the real world again. And that to some extent, this is one of the inputs driving this business output that we're talking about.”

“Like my, uh, point of view is that it's almost impossible to communicate a three-dimensional or four-dimensional experience on a two-dimensional screen.”

“biological age is like the clout score of your body.”