The good part starts at 11:11.

“It's like the opposite of lean startup. Like lean startup is like, yeah, you like see if you get something working and then you double down”

“if AI works overall, plenty of other people want all our compute. So they're willing to pay.”

“the promise of AGI is that we're going to spend almost none of the time doing things that we have to do”

“We had many checkpoints that were like green on 80 of the rows, but red on 20 of the rows. We knew that wasn't good enough.”

“I remembered like the internet rewards risk and it rewards like surprise and like things that like people don't expect.”

“The real question for AI is where can you apply that loop model in a way that you can make a better prediction or better idea, each loop”

“my theory on this, is that they're overcomplicating the task massively by trying to train the model to behave based on an 80-page ethical system that they've codified.”

“The modern quantum computers run at a clock speed of hundreds of hertz or thousands of hertz, not at the gigahertz that we're used to.”

“Pfizer would sell you a drug for a million bucks that was worth a billion dollars. They didn't care.”

“I didn't realize how petrified I was of not being seen as a great trader. What a stupid thing.”

“as context expands, and as individual user has a lot of personal context, the context window increase dominates the batch size.”

“You actually need to have a lot of depth in the users you have. You still need a lot of users.”

“I don't want an AI that just spits out slop that makes a vibe-coded clone.”

“And distillation is a way to effectively spend more money to get what is compute equivalent. Like outputs from a language model are a function of compute.”

“but fundamentally, like AI progress is about data and about writing better software to serve your models.”

“All the scaling laws are exponential increase in compute for linear increase in performance.”

“every great product, every great leader, every great decision comes from understanding that trade-off and realizing what does not matter.”

“The hard problem to solve is control, is once you're in the air, can you stay in the air?”

“without the vision of a goal, a man cannot hope to manage his own life, let alone the life of others.”

“it took a few hours to like outperform the two years of work which we'd done as with human in the loop”

“the pursuit and optimization of a very simple, easy to describe quantities can often give rise to incredibly hard to describe complexity and emergence.”

“I believe that the approach of monitoring the human language that these systems use is going to be increasingly difficult as agents get more powerful.”

“They very rarely, at least in my experience, have come from design sprints or times where you set aside on a calendar”

“What I found is the opposite. The team usually knows before the leader knows.”

“you're looking at people's eyes. And like, that is the metric. It is not a DAO. It's not a DAO over a Mao.”

“if you have a system where the improvee, the things being improved, is also the thing that's doing the improvement, you now have a new way to think about it.”

“we have one of our engineers that we named him the Viceroy of Variance.”

“presentations fail from too much information, not too little information. And I think that's true with books too.”

“What that means is that you just point your web browser at one malicious website, and the attacker will own your entire Windows PC.”

“the benefit of pooling is we get to essentially load balance across these different customer types”

“semiconductor manufacturing was invented to make one kind of machine, right? It's the semiconductor transistor and circuits built around that.”

“The companies that have been most successful that I've been involved with are ones where the pivot foot was about the technology, not about necessarily market understanding”

“The other thing I learned was when you just vibe code something for yourself, you have to maintain it.”

“I think about that as how many people were in a position to learn the thing that you did.”

“there will be certain points in reasoning that the model was initially sort of at a 50% chance of this answer and 30% chance of this answer.”

“There'll be some form of a blowup, a pretty high-profile one where people lose a whole bunch of money. and that will cause a reset.”

“what served you can also screw you. The dose makes the poison.”

“There's no announced plan for new revenue. Right now, it's shrink. It's cut costs. It's when we mush these two companies together, we're going to save $6 billion over the next three years.”

“If you had a thousand dollars in your pocket and someone took out five and ran away, you wouldn't then say, Hey, here's the rest.”

“I know to my core, the person I'm going to show you right now is the person I'm going to become 12 to 24 months from now.”

“you follow up forever until people buy, die, or they bought something and they don't want to, they don't want to work with you”

“if the pre-trained model has already memorized some of this data and then you try to do reinforcement learning on that, if it already knows the answer”

“magnesium diboride sat on people's shelves as a precursor for decades before then, and BCS theory had been invented back in 1957.”

“full autonomy is a non goal. It's sort of in the service that we use automation in the service of achieving the goals on the data.”

“So the attempt to try and centralize and manage the control of the software, I think is a fool's errand.”

“although this agreement was entered into voluntarily the governance that follows from it is not voluntary it is now day fast it's fiduciary”

“So unemployment is still near record lows at 4.1%, even though labor participation rose 0.2% to 61.6”

“I think the majority of actions on the internet will be taken by agents. Models are going to become more cheaper and faster at rates that are, you know, quite incredible.”

“we will pay our plugins that are popular and are seeing a lot of usage. They're going to get part of the revenue share as well.”

“The fundamental difference in the way that we've built dots is the harness does not run on the machine.”

“Where I get off the train is I think takeover is a lot harder than maybe the literal statement suggests.”

“When we see those indicators, we should probably pause and try to figure out what went down before going forward again.”

“they actually had synchronous classification or monitoring as a to-do item, but they didn't actually get to that to-do item”

“there's all the people in the org who don't really. We don't really want to maintain their own cause and then the people who do want to do it like eventually you kind of get tired of it”

“Builders want to be exposed to the complexity. Because that's everything you want to do is going in there and tinkering and turning knobs and whatnot.”

“I really don't see like, us, or the industry, repeating the chaos of like, every worker gets their coworker, and we're all gonna like double our company size, you know?”

“the church does not support machine consciousness. The priority is what he calls safeguarding the human.”

“the emotional sense in the room was like there's just such a weight.”

“you have to decompose their number. And so, okay, you did 2 million. That's great. How much of it was inbound? How much of it was outbound?”

“I can design the thing in three months, but I can't actually get it into real silicon at scale for nine months.”

“And they were open AI and the 01 was batting a 10 out of 10, which was amazing to me.”

“So the key thing is I need to call you right when you're taking the blood pressure measurement because now you're in that mode.”

“And then for the last 10 years, we've spent about 40 trillion on health care. We have not moved our life expectancy an inch.”

“What's really hard is that the customers believe or validate that your robot creates enough value that they want to deploy your robots across many of their sites.”

“whenever you're predicting abstract representation of the world you're encoding your own bias into what's important to predict”

“I think data quality in robotics is a non-stationary problem in the sense that what is high-quality data for your model changes over time.”

“I have a bribeable chief of staff. Like I have a corrupt butler. Is that what you want? That is a great term for it, a corrupt butler.”

“What I know about consumers is that the first time it doesn't work, they just walk away. They have no incentive to try again to overcome frustration or brokenness.”

“if you have failure recovery, if you have whatever it is that's interrupting your work, That's part of the issue.”

“if we projected inference to be 2% or 5% of the market, even if that specialized chip is, let's say, 2x faster, it might not make sense, right?”

“the demand for accelerated computers going up, but the demand for CPU and networking and storage, the traditional data center CPU, et cetera, is also going through the roof at the same time.”

“the glitch allowed us to get what one entrepreneur, Paul Alstrom, who wrote the book Nail It and Scale It, he calls it your beachhead.”

“Give me five seconds and then I get five minutes. They're okay with that. But I've earned it. Yes, because I've earned the attention.”

“for movies and video games, which are two of the kind of like biggest, you know, parts of the American economy, they spend more money on marketing than they do on actually making the product”

“So I think there's going to be potentially a 10 to 100x increase in volume and the number of queries that are going to hit up against databases.”

“they were able to get it up to like 99.5%, which is insane. I think that is, I think that's a viable thing now if you invest the time”

“the LLMs can get you maybe 85% of the way there and tuning certain things.”

“You can kind of think about it as for every generation of model, there's a 4X multiplier in compute, roughly.”

“I liken it to we start typically off renting our apartments, but then as you grow up, you want to own a house for various reasons, even if it's a bit of a headache.”

“I have not seen a straight shot sprint to an ownership market, but you have to go through the rental stage to become an owner.”

“You've got to break stars. You're going to launch stars. You need stars to put it behind the paywall so that people buy the paywall.”

“Egon goes and says, you're not going to remember this $500 million. $400. $400 million.”

“We tell ourselves stories in our heads that we then believe. And those are not real stories. Those are just stories. They're just not reality.”

“Anytime Greptile gives any sort of feedback or a score under five out of five, the grep loop fires. It takes in the feedback. It writes the code. It pushes to GitHub.”

“it forces the agent to prove that this bug actually happens. Right. Because sometimes it just might have been user error.”

“planning for me isn't the agent writing the spec and me reviewing the spec and me saying, go build it.”

“This is different because we're getting this capability before we've really characterized the shape of the threat.”

“the doom scenario becomes like way more human enabled if you are able to amass capital pretty quickly as a rogue agent.”

“for every breakthrough in math, you should assume there's a bunch of other breakthroughs in other fields that have more commercial applications”

“They took captured enemy weapons. They learned to fix conquistador swords to the end of long wooden poles, improvising pike tactics.”

“the force that marches on Tenochtitlan, the total numbers, the conquistadors were maybe one-half of one percent in raw numbers of that force.”

“when we train a protein language model, we train on metagenomic sequences, which are not the highest quality data.”

“if you think about live deployments at S&P 500 companies, that's at 69 percent.”

“how much incremental demand can I get from an AI agent? How many more orders can they bring me?”

“I think one thing I always try to encourage people to do, you know, whether you're a pstack user or not is to actually spend the time to sort of observe the bot or agent doing”

“we are creating bots as more persistent named agents that have their own identity, memory system, have their own tools.”

“We found that if someone is politically active as a teenager, but CZs to be in their early 20s that tends to correlate highly with someone's drive”

“I find that generally people lose motivation as they gain experience.”

“Every time someone recommends that we should add complexity, whether it's a step in a process, a feature, a team, head count to a team, you name it”

“the most at-scale workload, so the largest workloads where you're spending the most money, is actually where you're going to want to go and post-train first, right?”

“it is surprisingly easy to hill climb on non-verifiable domains if you turn them into some proxy verifiable domain.”

“I basically think what's blocking continual learning is basically extremely data-efficient training from sparse rewards, which is an unsolved problem.”

“They beat placebo in a statistically significant margin, but not necessarily in a clinically significant margin.”

“It's becoming such that like knowing what to prompt and knowing what to build is so much more difficult and valuable than just knowing, you know, the CTO being able to code the thing.”

“We pulled the data recently, and just in the last two years, YC has funded more than a dozen companies that are each making more than $10 million a year”

“Intuition is actually the most valuable when there isn't direct feedback because very many of the most important choices that we had to make”

“choosing the right of valid solutions is actually the hard part, not finding a right solution.”

“Here's the thing that LLMs and machines cannot do. Machines can't take responsibility.”

“my playbook was, err on the side of being just extremely picky, take a decent amount of time, start early”

“usually it's like one giant bottleneck. And I think you should at that point identify what that giant bottleneck is and then focus the whole company”

“to get that magical 10x, 100x, in a few rare cases, 1,000x productivity boost, you have to interact with the agents directly.”

“it might be the case that as we scale up RL, the extent to which you would have to clean up your environments is just going to be totally impossible.”

“reaching the championship and staying there reaching it is one thing but staying a champion that's totally different it's incredibly tough”

“If you can find something that's messed up and easy to un-mess up, booyah, there's your money. There's your opportunity to make a lot of money.”

“The biggest lever, the biggest component is the differential between what I raise capital at due to my relationships with mostly institutional investors”

“I like to ask them, if this was your money, would you buy this company?”

“all it had to do was convince a majority of the civilian population of the loyal states that it wasn't worth the cost to make these rebel states come back into the Union.”

“the biggest source of confidence is success. So if you get in the habit of making $100, well, next we make $1,000.”

“the pain of failure hurts more than success feels good. So what happens is we tend to index toward our failures, right? We bring the pain of the past failures to the moment.”

“We spend our adult lives undoing the debris of childhood. So just pause there for a minute and think about your life.”

“the math you get to is it's about $5 billion per gigawatt of CapEx to put these in space.”

“Frontier might be 90% of the economic value. Open source might be 80% of tokens.”

“we're spending $1.5 trillion of CapEx on $300 billion of inference revenue.”

“I think it is the absolute, the pinnacle of arrogance to think that what we can do, given the understanding that we have from what we've measured now”

“So that means at that rate, with that facility, it would take a billion years running with very little downtime to make a single gram of antimatter.”

“Both accelerators could either find it or definitively rule out the predictions of simple Higgs theory. 100% guaranteed.”

“LLMs do not have the ability to predict the consequences of their actions. and they do not have any planning abilities because inference is by predicting the next token”

“all of your information diet will be mediated by ai assistants”

“LLMs are good programmers they're not software architects they're not computer scientists right but they can program for us”

“They were about to find out just how powerful that technicality was. Or Jack was about to test it, shall we say.”

“If you invested $100,000 at age 25, and you got 7% market returns for 40 years, you'd end up with $1.5 million.”

“So when all is said and done, the Vanguard 500 returns a total of 126% net after fees for the 10-year period, while the hedge fund portfolio returns just 36”

“it is the greatest real estate scam in history”

“They would be traders in, say, an English port, kind of looking around, they'd get everyone's schedule, then they would sail away and come back as Vikings.”

“So he turns to one of his guards and says, you kissed the foot.”