Code

“The only thing that mattered for the last two years is LLMs do code, and code is a half a trillion dollar industry.”

“Anything that can be reduced to code will be done by it.”

“We've passed the point where humans will read every line of code. That is never happening again.”

“There isn't going to be less code. The code just becomes less valuable throw away.”

“There are so many companies now where software engineers have not had to write a single line of code in say more than six months, which is I mean, quite unprecedented, right?”

“Writing the code isn't the hard part. Writing the hard part is knowing what the code should be doing. It's seeing the problem and knowing the right way to do it. It's the opinion.”

“What I'm seeing is that code is becoming almost free. It's, it, the marginal cost is approaching zero.”

“Like with code or with math, it's relatively more binary, easy to verify. But how do you verify the quality of a board deck? Yeah. It's a, you have to be, you have to have like a good rubric based evaluator.”

“All of the difficulty in energetic engineering today is when people try to do things that should be in markdown in code and it fails because code is brittle.”

“The modern forms of leverage are code, media, and capital. Zero marginal cost and uncapped and accesses this sort of global marketplace.”

“I think the only way that you can actually get to understanding agent loops is if you build them in the material that they're made of, which is currently code.”

“Increasingly our designers are shipping code, which is really wild.”

“I don't think the LMS are going to naturally always deliver beautifully crafted, elegant, and yet scientifically correct code. And so you'll still need some degree of taste, even so much to converse with the LM to steer towards the right outcome. And for that,…”

“But it's a lot easier to change with code, as opposed to a model that is Learned primarily on data of real world and virtual worlds that are, I guess, like for example, junior, like there's actually trained on a lot of real world data and a lot of virtual gami…”

“Like code is easier than almost every other problem that you pose these agents because code is obviously right or wrong. Like it produces code. You run the code. Either it works or it doesn't work. There might be a few subtle hidden hidden bugs, but generally …”

“We're all moving in the direction where Our code is mediated through AI at some point.”

“And today, probably 95% of the code that I produce, I didn't type it myself.”

“I mean, we're like, we're not writing code by hand anymore. That's over. That's done.”

“Code used to be the moat. I think by now, maybe by the end of this, you understand that distribution is the new moat, right? It's the most important thing is distribution. You know, it's scarce, right? AI can't build it. Product is really important, but it, yo…”

“And now it's a hundred percent of code will be soon written by AI.”

“I am trying to get to a world where I'm not writing code.”

“I feel like I won't be like a self, Self-actualized software engineer in the future, if I'm too precious about that artifact, which used to be so central to me.”

“And then the other thing is I think most companies in Silicon Valley won't write code by hand.”

“And the very obvious statement is every single line of code that's generated has a marginal cost of zero, so therefore software, the lines of code, is no longer a moat. That's probably overstated.”

“But no matter what, there's going to be 10 to a hundred times more code. So I think you can be very long engineering right now as just a, you know, purely on the dimension of software is going to become increasingly more important. Once agents are, you know, t…”

“It's already probably the case that the majority of code being written is written by agents.”

“Just being in code, I think, is very linear. You might be running fast, but you might be running towards the wrong direction.”

“What's the layer that we are adding on top of code where people don't need to worry about code anymore. And at that point, it was like a year ago that we chatted, and it feels like this is the layer Is the agent conversation of what it is, what it's thinking a…”

“There's some hope that by training enough on code, the machine essentially like infers these kinds of structural cues.”

“First, 80% of code is getting written by AI. Now for large companies that have large legacy data code bases, it's a little harder, but like you all are seeing that a lot of the coding is written by AI”

“Law and healthcare, they're still a little too squishy, a little too soft. It's unlike math or code.”

“And I think of knowledge work as being on a similar sort of exponential as code has been on, but just time shifted, right?”

“I've written more code in the last one year than I have in the last 10 years, because code is now essentially architecture and English or whatever language that you have.”

“From what I've seen, you kind of reach a point where eventually you kind of have to look at the code and it's going to make your life a lot easier as you start to figure out how the code works. You don't need to be an expert but you can start to kind of tune t…”

“For math, it's a bit harder to over optimize, I think. Unless you have tools and the model learns to search and cheat instead of learning math, which I'm sure somebody could see that out in the world, which is like, oh, I'll just find the, you're training. It'…”

“I think it'll help the administrative side of things a lot, code things, where we literally can stick a drawing, ask AI to look at it for code.”

“One surprising thing is, we're used to thinking of code as this really valuable artifact because it's so hard to create, but because the cost of software engineering is going down, code is much less of a valuable artifact as it used to.”

“When we showed it to end users, a lot of times when people see code, they're like, oh, I'm not a programmer. This tool is not for me. It was like a visual turnoff to them.”

“One of the reasons that code is growing so quickly is because you have this really great feedback loop of success or failure, right? It's, you know, in, in, in healthcare or in law or something like that, you know, it's a lot harder to say, you know, whether y…”

“Capability in code is really core to performance on any LLM and like loosely, the better any LLM is at code, the better it is at any downstream task, even that's like writing poetry.”

“Code is one of the very few things, especially at the time that you could actually validate. And so you could have that agentic loop where the LM is generating the output and you're actually, you know, verifying in ground truth, the quality of that output.”

“Code has this whole automated feedback loop, right, where you can run the code, and that is the kind of automated feedback that really feeds into the RL, which makes these models so great at coding.”

“Code is this weird form of work where sometimes really the next five, 10, 20, 30 minutes of your work is entirely predictable from looking over your shoulder. And I would contrast this with writing.”

“For things like math or soon code that are verifiable, they will get solved very quickly.”

“Code is one of the first ones because it's so nicely automatable. There's a clean feedback loop and all of these things that you can really train models to get really good at code.”

“And it seems like everything that's quote unquote agentic right now is basically doing that, you know, across customer success and code. And you go through every domain that has like a specialized Approach and they always have this sort of orchestration layer …”

“That is really changing now where designers can essentially through software, through code can actually be really close To building the object. Right. Versus kind of like, you know, back in the day, maybe it's furniture or like other objects. And so I think th…”

“At least in my experience, rag from language to code doesn't work super well.”

“What doesn't a product manager do? Write code. Right? So that's gonna change.”

“Rule of thumb, I would assume that you'll throw away 90% of all the code you write. So your goal at that stage is just to identify the 10% of what you're building that's actually valuable as fast as possible.”

“I believe that code is usually a certainty. Code is something you can hire someone to do because building that next feature does not have any risk. Versus marketing and sales usually have a ton of the risk, as well as product, like deciding what to build, How …”

“Sometimes I say that like the LLMs aren't [4137] Great at planning yet. [4138] So we can help them plan by telling them how to plan and code. [4140] Cause that's very explicit and that's a good way of communicating how they should plan and stuff like that.”

“I believe so. Yeah. I said actually two years ago at a conference that my prediction back then was 80% of code is going to be written by AI in five years. So I guess I have three years to go to, for that to become true.”

“Code is probably the only domain that actually has a lot of tokens.”

“I think in a, you know, a, you know, a review of, you know, your VP of engineering, you should be looking at the code that that team shipped.”

“It's a lot more difficult to get to the top selling a physical product than it is to get to the top using code.”

“GitHub CodePilot is now writing 50% of code, and that's every other line of code to put that in perspective.”

“And as you do things that you want to do more robustly and you want to do in a more repeatable way, then you want to move it more towards code, right? And so To the extent that you've never seen this task before, maybe you should be doing it in this more kind …”

“One of the reasons why we work on code is that there are objective answers to a lot of these questions, either the test pass or they don't, either the function is correct or it isn't. Those kinds of things are much easier to evaluate.”

“Code is not complicated. You know, I don't, I can't write it well, but I can understand it. You can look at it. It's just a logical system.”