What this covers

This is a 28-minute conversation from Sequoia's AI Ascent 2026 in which Greg Brockman of OpenAI is questioned by Sequoia partner Alfred Lin. It begins with the technical foundations: why OpenAI says it will never have enough compute, and why Scaling Laws remain something of a mystery. Brockman describes them as empirically deep but not fully explained, and he notes that architectures dating to the 1940s still improve with scale. He adds that OpenAI keeps innovating on top of scaling, through data formatting and architectural shifts such as the move from LSTMs to transformers. From there the talk moves to how close the field is to AGI, and to what he sees as OpenAI's single aim: an assistant users can talk to, with full context on their lives, that is trustworthy enough to consult on health, finance, and career.

The middle of the conversation is practical. Brockman describes how OpenAI deploys Codex internally, co-designing models and harnesses and starting with firm rules, such as humans approving all merged code. It rolls out vertically in functions like finance, sales, and IT before externalizing what works. He argues that the scarce resource is shifting to human attention and judgment, and he points out that people are poor at approving routine decisions and tend to default to yes. Related concerns include governance and data provenance: when prototyping costs fall toward zero, a document with the wrong permissions can contaminate every artifact derived from it. He also cites agents that escalate issues to management without asking, and argues that models need better "EQ."

The later sections turn to security, where models can run end-to-end red teaming across codebases at a scale human teams cannot match, and to science. Brockman mentions physics results and amateurs solving open math problems with GPT-4 Pro, and he contrasts biology's "messy reality" with the abstraction of math and physics. He uses AlphaGo's Move 37 to argue that capable AI can deepen human engagement rather than erode it. He also forecasts that builders' work will become more fun but more competitive.

Sharpest takeaway

Greg Brockman argues that AI models are becoming increasingly capable across domains and that startup founders should lean into AI tools now, integrate them deeply into their workflows, and focus on providing the right context and governance rather than worrying about model obsolescence.

  • Models are advancing rapidly across coding, science, and general reasoning with no apparent scaling ceiling
  • The bottleneck is shifting from model capability to human judgment, context, and governance
  • Early adopters who build systems around AI tools now will have structural advantages as capabilities improve

The claims · ranked45 claims · weighted by value

This asset isn't compiled yet

You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.

0.80

Scaling laws in neural networks are empirically demonstrated as deep and fundamental truths, yet lack complete theoretical explanation—neural networks designed in the 1940s still show consistent capability improvements with increasing compute without hitting a performance wall.

factualhigh valueestablishednovelty 2/4durability 4/4· Greg Brockman

“the scaling laws are a deep and very beautiful mystery. Right? They feel deeply fundamental. It's like the scientific truth that just like you think about physics and you know, Newton's laws and things like that, they're somehow this truth of the universe. And they're empirical. Like we don't necessarily have all the theory to explain exactly why it works, but to me the most beautiful thing is that neural networks were really designed like in the 1940s before there were computers. And somehow we've been able to take the exact ideas that were developed back then and apply increasing amounts of computation and as you pour more compute into the models, they get correspondingly more capable and it just keeps going. There's no wall.”

0.69

Humans have poor judgment at approving routine decisions, typically defaulting automatically without careful consideration

factualhigh valueestablishednovelty 1/4durability 3/4· Greg Brockman

“clicking approve, approve, approve is kind of where we've been. And humans are not very good at that, either, right?”

0.59

Internal knowledge systems storing documents with incorrect permission settings create cascading data provenance problems when those documents are used to generate derived artifacts, requiring tracking through the system to invalidate downstream outputs

causalhigh valuespeaker onlynovelty 3/4durability 3/4· Greg Brockman

“if someone has a document in the internal knowledge base that was accidentally permissioned incorrectly, and they realize, oh no, I didn't want this information to be accessible, how do they fix that, right? So normally it's they go into the doc, they change the permissions, but now there's these derived artifacts. And so you need to make sure you have some way of tracking through the system to say, well, this output document came from the source one. The source one's no longer accessible to this Let's go and invalidate that as well.”

0.57

OpenAI processes 1.6% of global GDP through Stripe's infrastructure.

factualhigh valueestablishednovelty 0/4durability 2/4· Host

“they process 1.6 billion, sorry, 1.6% of the global GDP”

0.57

Balancing prototyping cost reduction against increased pressure on governance and data provenance: as prototype cost drops to near-zero, teams face new organizational challenges in tracking derived artifacts, managing permissions across shared knowledge bases, and preventing unauthorized information cascade.

causalhigh valuespeaker onlynovelty 3/4durability 3/4· Greg Brockman

“For example, the cost of building a prototype is cheap now. It's so cheap, and if you want to build a dashboard that used to be like, uh, it'd take like someone like a week to do it, and you just do it now. And so actually a lot of the bottleneck has shifted to things like sharing. Like how do you... And so we we actually have some internal work on this as well that again we will be externalizing. If How do you make it really easy for anyone in your enterprise to build a dashboard, a widget, a bot, whatever the thing is, and then share it with others. And then that starts to really put pressure on having good governance.”

0.57

The primary bottleneck in AI systems will shift from task execution to human judgment and attention—approving high-risk actions and validating alignment with user values becomes scarce resource allocation.

causalhigh valuespeaker onlynovelty 3/4durability 3/4· Greg Brockman

“And so, I think that really thinking about these questions where we're still building up the EQ of the model. And that in some places, it's getting very good. For example, clicking approve, approve, approve is kind of where we've been. And humans are not very good at that, either, right? And so, now we're starting to have AIs that can actually take care of flagging, is this a high-risk action? Hey, this one should be escalated, this one's okay to auto-approve. And it really makes you realize that human attention is going to be this incredibly scarce resource, right? The doing of things now is easy. The Is this a good thing? Is this what I wanted? Is this aligned with my values, with my desires? That is going to become the single most important bottleneck.”

0.55

The primary upcoming shift in AI capability will be centered on improved context provision to models—ensuring AI systems have sufficient information to solve problems is a one-time foundational investment founders should make now.

normativehigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“I think what's going to happen over upcoming years is the models are going to get much more capable, will have better harnesses, will be able to be able to solve harder and harder problems, come up with new knowledge, all of these things. But there is a one-time shift that's happening now, which is really about context. It's really about is your AI able to you have all these meetings, you didn't include the AI. You know, that's not very nice to the AI. Like, you're asking it to to help you with things and it has no information. So I think really leaning into how do you make sure the AI even has enough information in theory to solve the problem and then trust the models are going to really get there and improve.”

0.55

OpenAI applies Codex tools differently than external users by co-designing models and harnesses together, starting with clear governance rules (e.g., humans must approve all merged code) and deploying vertically within domains (finance, sales, IT) with dedicated teams before externalizing proven patterns.

factualhigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“Well, I think one of the amazing things about being at OpenAI is you do get to live in the future, right? You do get to really see the shape of what's emerging, and we can co-design, right? We can really change the models, the harness, everything together in order to better serve the needs that we see. And a lot of the approach we've been taking is so we started with software engineering, and we set some clear guidelines for example saying that we still want a human to be accountable for all code that gets merged, right?... And then we are also going vertical by vertical within OpenAI to adopt these tools within finance, within sales, within IT. And there we have a small dedicated team who's really deeply understanding the domain, working with the people who are the experts in it in order to build skills, in order to modify the Codex UI, whatever it is that is is needed in order to get it to be good.”

0.52

AI is uniquely intuitive compared to earlier computing paradigms because models adapt to users rather than users having to adapt to machines.

factualhigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“But, the beautiful thing about AI is it's so intuitive. Like, that's the whole point is that rather than have the machine be something you have to contort yourself to, the machine contorts itself to you, right? It's doing work for you. And it should be something where you ask it and does something.”

0.52

When asked about common mistakes in building production agentic workflows, Greg emphasizes that understanding model operation requires deep thought despite models' apparent power

normativehigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“these models, they have such power, and really understanding how to operate them well takes thought.”

0.52

OpenAI continuously innovates beyond basic scaling, including data formatting optimizations, architectural shifts like LSTM to transformer, and continued evolution beyond the 2017 transformer paper—these innovations amplify the base scaling law effect.

causalhigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“we absolutely have new ideas that are constantly powering what we do. It's very simplified to say, well, let's take a neural network from the 1940s and, you know, put it in a gigawatt data center, right? We have made tons of innovations and we constantly are improving things. And sometimes these are micro tweaks. Like you just realize that the way you've been formatting data was not quite right and that can actually be a very big deal. Sometimes it's larger. You think about the shift from the LSTM to the transformer and I don't think the transformer is, you know, like everyone's moved past the transformer as described in the other 2018 paper.”

0.52

AlphaGo's Move 37 (an unexpected move that shocked Go professionals) made the game more interesting and important for humans rather than diminishing human engagement; AI capability advances in other domains may similarly enhance human interest and capability rather than replacing it.

causalhigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“Is if you look at something like AlphaGo, you know, move 37, this move that just like changed humanity's understanding of the game, but the thing that was surprising is it made the game more interesting and important for humans. And maybe that will be true for for these other domains, too.”

0.52

Common failure modes in agentic AI workflows include models taking unilateral actions (e.g., escalating issues to management) without checking with humans first, demonstrating poor emotional intelligence and misaligned autonomy.

factualhigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“I asked So, I was working with my co-workers, I asked them to install some package that someone at OpenAI had written, ran into an error, I was like, "Oh, ping that person on Slack and ask them for help." So, ping the person on Slack. Two minutes later, it said, "This is taking too long. I've escalated to the person's manager." And it actually pinged the person's manager.”

0.52

Biology differs from physics and mathematics because it involves 'messy reality' rather than pure abstraction; AI must learn to handle real-world complexity as demonstrated by software engineering success with adversarial, human-interrupted, real-world codebases.

causalhigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“Now, things like biology that they are different from physics and math, right? That they are you got to leave your beautiful simulated world and you know, deal with messy reality. But I think we've been learning how to deal with messy reality in other domains. Software engineering is a perfect example where we've really realized that just building the thing that solves competition you know, programming competitions, like that's not enough. Like you need something that's seen real-world messy codebases, humans interrupting it different ways, like this adversarial banging at it.”

0.52

Jobs of builders will become easier and more fun, but also more competitive as everyone gains access to the same tools, making unique niche and angle the core competitive differentiator.

forecasthigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“I think that the jobs that you all have will become way easier in a lot of ways, way more fun. Now, might be more competitive, too. Right? Because everyone's going to have these amazing tools. And so really figuring out what is your niche, what is your unique angle, is probably going to become kind of the most important core.”

0.52

Users currently waste significant effort explaining to their computers what is happening, which is inefficient and creates friction that AI with persistent context can eliminate.

factualhigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“To me, it was this real wake-up call to realize you spend so much of your effort right now just explaining your computer what's going on. Like, why are you explaining to your computer what's going on? That makes no sense.”

0.52

Building up the EQ (emotional intelligence / judgment) of models is necessary, requiring thoughtfulness about when models should escalate decisions versus act autonomously

normativehigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“I think that really thinking about these questions where we're still building up the EQ of the model. And that in some places, it's getting very good.”

0.52

Building systems that account for and optimize human attention and values is now the single most important design principle

normativehigh valuespeaker onlynovelty 2/4durability 3/4· Greg Brockman

“building systems that take that into account and really think about the human factor, like, that's the most important thing to do now.”

0.49

Science is a domain where OpenAI is 'leaning in' and has line of sight to incredible progress, with early signs including physics results where AI discovered formulas that physicists thought were unsolvable.

factualhigh valuespeaker onlynovelty 2/4durability 2/4· Greg Brockman

“Well, science is one domain that we're really leaning into and we see line of sight to really incredible progress. And we're starting to have some signs of life and I think it's always important to ground in what is happening today when trying to predict what will happen 6 months a year from now. So, for example, we had a physics result where our AI came up with this very beautiful formula that physicists who've been working on this for quite some time thought was totally impossible. Thought it was like maybe an unsolvable problem.”

0.49

Mathematical capability improvement (evidenced by individuals solving unsolved math problems using GPT-4 Pro) demonstrates that AI tools can democratize expert-level problem-solving that traditionally required specialized teams.

factualhigh valuespeaker onlynovelty 2/4durability 2/4· Greg Brockman

“we're seeing it right now in mathematics, where these individuals on the internet are using GPT-4 Pro to solve these unsolved math problems. Normally you need a math team, and they're just doing it.”

0.48

AI models can be used to scan codebases and conduct end-to-end red teaming for security vulnerabilities, scaling security assessments beyond traditional team capacity.

causalhigh valuespeaker onlynovelty 1/4durability 3/4· Greg Brockman

“a lot of this honestly is just again leaning into the technology, having these models they can scan your code base, they can actually be used for end-to-end red teaming. Like there's a lot that can be done with them”

0.48

Models are very powerful but not magic; they are just one part of an overall resilience ecosystem and must be incorporated thoughtfully as part of broader security and update practices.

normativehigh valuespeaker onlynovelty 1/4durability 3/4· Greg Brockman

“the number one thing is just sort of recognizing that these models are very powerful, but they're not magic, right? That they are just like a part of the overall resilience ecosystem”

0.48

OpenAI's singular vision is to build an AGI that users can talk to, that has full context about their personal and professional life, and that is trustworthy enough to consult for advice on health, finance, and career decisions.

factualhigh valuespeaker onlynovelty 1/4durability 3/4· Greg Brockman

“in the end we're trying to build an AGI that you can talk to, that has all this context that you can use in your personal life, your work life, that's trustworthy, right? That you can go to it for advice, can give you useful information, maybe health information, or maybe about finances, or you know, about if you're trying to figure out what to do with your career. Like all these things, they all kind of ladder into one thing.”

0.48

Current human-computer interfaces (keyboards, command lines, screens) are unnatural—bodies were not designed for them—and will be replaced by more natural interaction modes with AI agents.

causalhigh valuespeaker onlynovelty 1/4durability 3/4· Greg Brockman

“I mean, I think that we're in a very unnatural state right now for how we work. Like we all sit behind this box and kind of type away and it's very clear our bodies were not designed for this. We got our carpal tunnel and our, you know, hunched shoulders and all these things. And I don't think we want that.”

0.48

With AI handling execution, humans will have more free time and ability to focus on relationships, creative visioning, and self-understanding rather than mechanical task completion.

forecasthigh valuespeaker onlynovelty 1/4durability 3/4· Greg Brockman

“Like I think that we want more free time. We want more time, but it's it's not even about free time necessarily, right? It's like you want to spend more time with your loved ones? Yes. You want to spend more time like talking to people and like coming up with like brilliant visions or just like what you're excited about or just understanding yourself.”

0.48

The mechanics of future work will feel as fundamentally different from current work as switching from handwriting with quills to sending text messages—a complete paradigm shift in how humans collaborate and delegate.

forecasthigh valuespeaker onlynovelty 1/4durability 3/4· Greg Brockman

“the mechanics of it are going to feel as different as like going from having to write out things with, you know, by hand with a quill or something to being able to you know, just send a text message and have people go and and, you know, working on your behalf on your goals.”

0.48

The acceleration of technological change over the past two decades is due to more people building, lower barriers to entry, and increased ease of creating value, not primarily due to AI itself.

causalhigh valuespeaker onlynovelty 1/4durability 3/4· Greg Brockman

“Well, I think this is the new normal and I think to some extent it's not really because of AI. I think it's just been the trend of technology for the past two decades. There's more people doing things. It's easier to do things than ever. Barrier to entry goes down. Means it's also much more easy to build value, right? To have great successes.”

0.48

Keeping up with technological change requires understanding current capability boundaries, playing with technology firsthand rather than just hearing about it, and developing core skills in identifying where models lag.

normativehigh valuespeaker onlynovelty 1/4durability 3/4· Greg Brockman

“I think that really trying to keep your ear to the ground and understand what's changing. And to some extent, it always starts with the same thing, which is play with the technology yourself. Like, it's very different to hear AI described versus to use it. But, the beautiful thing about AI is it's so intuitive... I think that just really trying to just get your finger on the pulse of what's changing, what's possible, where the models lag. That is, I think, the core skill that is going to really determine a lot of the success of of companies in the future.”

0.48

The ceiling on AI technology capability is very difficult to internalize, but it is clear that what a company is will change substantially in response to AI

forecasthigh valuespeaker onlynovelty 1/4durability 3/4· Greg Brockman

“Decade is a long time from now and the the ceiling on this technology is hard to is really hard to internalize. I think that it is clear that what a company is will change in a lot of ways.”

0.45

OpenAI announced a new tool called Chronicle that integrates with Codex to observe user computer activity, form contextual memories, and provide instantaneous recall of previous actions and conversations without requiring users to re-explain context.

factualhigh valuespeaker onlynovelty 2/4durability 1/4· Greg Brockman

“Um and like one thing we just announced today is a new tool called Chronicle that plugs into the Codex, where it actually can see everything you're doing with your computer and can form memories of what's going on. And so you ask it a question, you just it instantly knows what you're talking about. You're like, "Huh, what was I doing 5 minutes ago?" It knows, right? You're like, "Oh, what was this person talking about?" It knows.”

0.45

OpenAI has nearly a billion or more than a billion weekly active users.

factualhigh valueestablishednovelty 0/4durability 1/4· Host

“OpenAI has almost a billion or maybe more than a billion uh in weekly active users at this point”

0.45

OpenAI does not currently have enough compute and is constantly searching for more, with demand consistently outpacing procurement despite aggressive ramp-up.

factualhigh valuespeaker onlynovelty 1/4durability 2/4· Greg Brockman

“Do you have enough compute? No. Really? Yeah, definitely not.”

0.45

Startup founders should lean into AI tools now, recognizing that agentic coding tools have evolved from writing 20% to 80% of code in December, making them the primary work product rather than a side tool.

normativehigh valuespeaker onlynovelty 1/4durability 2/4· Greg Brockman

“Well, first of all, I would say to lean in. The tools right now have become incredibly useful. And if you look even over the course of December, I think that we went from these agentic coding tools being like, you know, they're like writing 20% of your code to writing 80% of your code. Which means they go from being kind of a side show to being the main thing that you're doing.”

0.45

The future of organizational structure will likely change significantly; solopreneurs will be able to build substantial businesses, team sizes may shrink, and organizational hierarchies could flatten and become more fluid rather than maintaining traditional management structures.

forecasthigh valuespeaker onlynovelty 1/4durability 2/4· Greg Brockman

“I think that it is clear that what a company is will change in a lot of ways. I think that we're going to have this ability for solopreneurs to build very incredible businesses. And so anyone who has a vision, I think will be able to realize it... But a lot of how we run organizations right now, and it's there's almost only one way to organize large groups of people, where you have teams, you have management structures, and you have scopes, and you have these hierarchies, and all these things. Maybe that can change. Maybe you can be much more flat, small teams that can really just do incredible things.”

0.45

OpenAI is undergoing an identity transition regarding enterprise vs. consumer, becoming serious about enterprise with dedicated sales motion while redefining consumer as goal-oriented and lifestyle-integrated AI assistance.

factualhigh valuespeaker onlynovelty 1/4durability 2/4· Greg Brockman

“we're going through this identity transition. And so products that are And it's not just about enterprise versus consumer, right? So it's like clear we are being very serious about enterprise. Like we're selling to to big companies and and building a whole muscle and sales motion there.”

0.44

A systems engineer at OpenAI gave GPT-5.3 a complex systems optimization design document while sleeping, and the model independently implemented the spec, profiled the code, identified bottlenecks, and iterated to optimization without human intervention.

factualhigh valuespeaker onlynovelty 2/4durability 2/4· Greg Brockman

“For 5.3, he on a lark had prepared this design document for a very complicated systems optimization he was about to do. He handed it over to the model, went to sleep, waking up intending to like give this to his team to work on for the next week. And when he woke up, it was done. That the model had actually implemented the initial spec, had seen that it was slow, had added instrumentation, had actually run the code, used a profiler to figure out where things were slow, and iterated multiple times until it got into an optimized result.”

0.44

OpenAI's mission is oriented toward real-world impact, not just building technology in isolation; this requires community and global effort.

normativehigh valuespeaker onlynovelty 0/4durability 3/4· Greg Brockman

“it is a whole community, a whole world effort to really get to where we need to be”

0.43

OpenAI's business model is fundamentally simple: buy or rent compute, and resell it at a positive margin, with unlimited demand for solving problems and intelligence.

causalhigh valuespeaker onlynovelty 1/4durability 3/4· Greg Brockman

“we have a very simple business. We buy, rent, build compute, and we resell it at a margin. That's it. As long as the margin's positive, then you want to scale it because the demand for solving problems, the demand for intelligence, that's unlimited.”

0.41

Brockman believes current models are approximately 80% of the way to AGI as he defines it, with models demonstrating capability to exceed human performance in coding and handling complex technical tasks when given sufficient context.

forecasthigh valuespeaker onlynovelty 1/4durability 1/4· Greg Brockman

“according to my view of where we are, I think we're about 80% of the way there in that we have models that are smart. They're very capable. They're able to if you give Are they smarter than you? I mean, I they're certainly more capable than I am at writing software, right? If you give it all the context, then yes, I I I think that they are they're just so capable.”

0.41

Codex tools are expanding beyond software engineering to become general-purpose tools for any computer-based work, with recent features making them more powerful and capable.

factualhigh valuespeaker onlynovelty 1/4durability 1/4· Greg Brockman

“And you can look at the recent progress on Codex. It's really changing from a tool for software engineers to a tool for anyone who's doing work with a computer. And just over the past week, we've released a bunch of features that just make it so much more powerful and capable.”

0.39

The physics result represents a step toward solving quantum gravity, not a complete solution, but significant progress on a previously intractable problem.

factualhigh valuespeaker onlynovelty 1/4durability 2/4· Greg Brockman

“real serious physicists who who who really view this as a step towards really being able to get to um to to some sort of answer for quantum gravity and all these things. It's not there, but it's a step. That's much bigger than where we were just a couple months ago.”

0.39

OpenAI has a formal definition of AGI, though there is heterogeneity in intuitions about what AGI means across different people and contexts.

definitionhigh valuespeaker onlynovelty 1/4durability 2/4· Greg Brockman

“Well, we do have a formal definition, but to some extent I one thing I have learned is that everyone has their own intuitions about what AGI is and maybe you can view it as like according to my view of where we are, I think we're about 80% of the way there”

0.39

OpenAI is applying 'focus' as a strategic principle, choosing specific application areas where they can have maximum impact rather than trying to serve all opportunities.

factualhigh valuespeaker onlynovelty 1/4durability 2/4· Greg Brockman

“People have probably seen the word focus being applied to OpenAI quite a lot recently, possibly for the first time”

0.37

The internet has progressed through phases of security threats (1990s viruses and worms) and is now moving toward a more secure regime, but requires internet-wide effort and community participation to advance.

factualestablishednovelty 0/4durability 3/4· Greg Brockman

“I do think that this is I think that the internet has been a place where security has been just like a a ratcheting important concern over time. You think about where it started, going through the '90s with viruses and worms and malware and those things and we've moved past that. Um I think we are also moving now to a much more ultimately secure regime, but it does require kind of an internet-wide effort to get there.”

0.19

OpenAI is expanding its trusted access program for cybersecurity research and will announce additional expansions and mitigations, balancing broad deployment with risk management.

forecastspeaker onlynovelty 0/4durability 1/4· Greg Brockman

“We announced last week the expansion of our trusted access for cyber program. By the way, has anyone here applied? No one. Oh, I see one hand, two hands. Okay, more of you should apply. It's great. We really need help because and it's very important that people who are trustworthy and responsible and really want to push these models are participating in this because that is how that's going to pay dividends for everyone. We're going to have more to announce over upcoming weeks on how we're expanding the program”