The Ledger
Every statement that passed quotation and attribution checks. Mix any filter with any other: certainty 1/5, debate potential 5/5, or both at once.
why aren't all 3,312 resolved? a statement only gets an assessment when the public record can support or contradict it. opinions and what-ifs never can, and 43 checkable ones are still open, waiting for their date. predictions held up or didn't; assertions are supported or contradicted. on every card: ▮▮▮▮▮ certainty · ▮▮▮▮▮ debate potential. speakers are clickable
Fedus: Self-improving software AI will not automatically generalize to domains like biology
“So rolling forward that software engineering self-improvement, I think you're going to have a system that can write complete repositories, identify bugs, refactor code, But it doesn't suddenly understand biology. Right? It's just like there's a domain gap ther…”
Fedus: Recursive self-improvement is happening in software and entering AI research
“So, In that domain, I think it's happening now-ish. And I think we'll see the same thing too for AI research. That's a slower outer loop because now the experiment isn't just checking some unit tests passing, but it's checking what was the scaling property? Di…”
Karpathy: Coding agent failures are user skill issues, not capability limits
“Even if they don't work, I think to a large extent, you feel like it's a skill issue. It's not that the capability is not there. It's that you just haven't found a way to string it together of what's available. Like I just don't, I didn't give good enough inst…”
Karpathy: Claude feels like a teammate, whereas Codex coding agent is dry
“I actually think Claude has a pretty good personality. It feels like a teammate and it's excited with you, et cetera. I would say for example, Codex is a lot more dry which is kind of interesting because in Chashi PT, Codex is like a lot more upbeat and highly…”
Karpathy: Software industry must reconfigure for agent customers instead of humans
“So I think the industry just has to reconfigure in so many ways that it's like the customer is not the human anymore. It's like agents who are acting on behalf of humans. And this refactoring will be, will probably be substantial in certain sense.”
Karpathy: Autonomous agent tool creation will be trivial within three years
“I kind of feel like this kind of stuff that I just talked about, this should be free, like in a year or two or three. There's no bi-coding involved. This is trivial. This is table stakes. This is like any AI, even the open source models, et cetera, can like do…”
Karpathy: Maximizing AI leverage requires removing humans from prompting loops entirely
“To get the most out of the tools that have become available now, you have to remove yourself as the bottleneck. You can't be there to prompt the next thing. You're, you need to take yourself outside
you have to arrange things such that they're completely auto…”
Karpathy: All frontier AI labs are pursuing recursive LLM self-improvement
“What I'm more interested in is, like, this idea of recursive self-improvement and to what extent you can actually have LLMs improving LLMs, because I think all the Frontier Labs, this is, like,
The thing for obvious reasons.
And they're all trying to recursive…”
Karpathy: A research organization can be defined as a set of markdown files
“A research organization is a set of markdown files that describe all the roles and how the whole thing connects.”
Karpathy: Smarter AI models do not automatically gain broad societal capabilities for free
“The story is that we're getting a lot of the intelligence and capabilities in all the domains of society, like, for free as we get better and better models, and it's not, like, exactly fundamentally what's going on, and there's some blind spots, and some thing…”
Karpathy: AI models will speciate into specialized smaller domain experts
“I do think that we will, we, I do think we should expect more speciation in the intelligences.”
Karpathy: Swarms of internet agents could outpace frontier AI labs
“A swarm of agents on the internet could collaborate to improve LLMs and could potentially even, like, run circles around Frontier Labs. Like, who knows, you know? Yeah, like, maybe that's even possible. Like, Frontier Labs have a huge amount of trusted compute…”
Karpathy: Cheaper software via AI will trigger Jevons paradox increasing demand
“So if the barrier comes down, then actually you have the Jevons paradox, which is, like, you know, you actually, the demand for software actually goes up. It's cheaper and there's more”
Karpathy: Frontier AI researchers are actively automating themselves out of jobs
“Even with other research, like, OpenAI or, you know Anthropic or these other labs, like, they're employing, what, like, a thousand something researchers, right? These researchers are basically, like, glorified auto, like, you know. They're, like, automating th…”
Karpathy: Frontier lab researchers face pressure over what they can say
“Like, if you're inside one of the frontier labs, like, there are certain things that you can't say and conversely, there are certain things that the organization wants you to say, and, you know, they're not gonna twist your arm, but you feel the pressure of, l…”
Karpathy: AI researchers lose technical judgment after leaving frontier labs
“And I think if you're outside of that frontier lab your judgment fundamentally will start to drift because you're not part of the, you know, what's coming down the line.”
Karpathy: Open source trails closed frontier AI models by 6 to 8 months
“So there may be, they're behind by like, what is the latest, maybe like eight months, six months, eight months kind of thing right now.”
Karpathy: Solely relying on closed AI models creates systemic risk
“I don't actually think it's like structurally, I think there's some systemic risk attached to just having intelligences that are closed, and that's like, that's it. And I think that that's a, you know, centralization has a very poor track record in my view”
Karpathy: Educators will soon teach AI agents instead of human students
“I feel like there's gonna be less of like explaining things directly to people. And it's gonna be more of just like, does the agent get it? And if the agent gets it, they'll do the explanation.”
Karpathy: Software documentation will shift from human HTML to agent markdown
“Instead of HTML documents for humans, you have markdown documents for agents, because if agents get it, then they can just explain all the different parts of it. So it's this redirection through agents, you know and that's like, so I think we're going to see a…”
Simon Last: Companies Must Redesign AI Software Harnesses Every Six Months
“I think this is honestly a really key thing and something that a lot of companies get wrong is just like doing one thing and then just like sticking with it. You really do have to keenly aware of what the current state of the models and the technology is, and …”
Simon Last: AI Coding Agents Can Build More Robust Software
“The upshot is I think if you do it well, you can be much more ambitious about what you're building. And also make it much more robust than you could have done with humans writing it. And then the flip side is if you do it badly, it's all slop.”
Simon Last: Current AI Tools Enable 100x to 1,000x Engineering Leverage
“You can be like a hundred or thousand x engineer for using the tools right now. I think the gap is much bigger. Like the minimum bar has not changed, but the maximum bar has extremely increased.”
Simon Last: AI Agents Now Write All Pull Requests at Notion
“And you know, all the pull requests are now written by agents. They're often like larger and more complex. That's like the worst part. But the better part is that they're often like a much better tested and we can demand sort of a much better testing for the t…”
Simon Last: AGI Will Essentially Be a Coding Agent
“I think of coding agents as, like, the kernel of AGI. AGI will be a coding agent.”
Simon Last: Notion Now Uses Cheaper Chinese Open-Source AI Models
“We actually just released one of them in our agent last week, and we're gonna do all four for sure. They're actually quite good, and they're way cheaper than the Frontier models.”
Tiwari: Hyperscaler AI CapEx Will Scale to Trillions Over Next Several Years
“And over the next several years you know, that scales to trillions of dollars, right?”
Tiwari: AI Compute Debt Uses Customer Cash Flows, Not GPUs, as Primary Collateral
“And I think what got missed was the GPUs themselves were actually like the second, second or tertiary level of collateral in those instruments. The primary collateral was the contracted cash flows from investment grade counterparties.”
Tiwari: Blackwell GPUs achieve 90 to 100x inference efficiency over Hopper
“And so for the hoppers, the H 100 or H 200 series of GPUs into the Blackwells there was a claim made that it could be 30 times more efficient. And I think the data from, you know, some analysis showed that it was 90 to a hundred times more efficient in terms o…”
Tiwari: There are no idle 'dark GPUs' in AI data centers
“And I think what you see here is I have, you know, you don't see any dark GPUs.”
Gil: AI engineering productivity boosts will be absorbed by infinite software demand
“I think that fundamentally you know, there's so much demand for software and there's so little supply of engineering and reality relative to that demand that as you add this enormous boost of productivity to software engineers it just gets soaked up. Right. Be…”
Guo: High-status engineering tasks can be surprisingly easy for AI agents
“One thing I've seen is that if you have An engineering identity that's based on that, like a value based ranking of difficulty or skill, like the specific types of engineering that are considered, you know, impressive or high status can actually be like less h…”
Guo: Unreviewed AI-generated code introduces fragility into production codebases
“If you can generate an enormous amount of code and no one is reading it, you don't know the quality of the code. Nobody deeply understands the code base and there's more fragility, right? It's like the slop problem, but instead of it being like vibe coding slo…”
Gil: Claims that AI agents make vendor purchasing decisions are overstated
“I think the one other thing people are bringing up that I don't quite buy is that agents are already making like big decisions for vendor purchases and things like that.”
Gil: Most Startups Have a 12-Month Peak Valuation Window Before Crashing
“Because often for companies, there's about a 12 month window where your company's the most valuable will ever be, and then it crashes out. For a very small handful of companies, the answer is you should never, ever, ever sell. For most companies, the answer is…”
Gil: Building a bundle is software's best defense against AI shifts
“The best way to defend against this is to build a bundle.”
Gil: AI compresses decade-long tech displacement cycles into two years
“And so the difference with AI, Is the velocity of change is so high that what normally would have taken a decade, and you'd have a normal decade-long displacement cycle, is now happening in a year or two. And that's really the reason that these things are so t…”
Scaringe: Fewer than five non-Chinese companies can solve neural net autonomy
“I think there's more than one, less than five companies outside of China that have The necessary ingredients to do this, the capital, the GPUs, the car park with, you know, enough vehicles generating enough data. I say more than one less than five. It's probab…”
Scaringe: Rule-based autonomy has a zero percent chance of competing
“A lot of the solutions that are more one dot O based and are sort of stuck in that framework, I think have a, like a truly a zero percent chance of progressing to be competitive with a neural net based approach.”
Scaringe: By 2030, car buyers will expect vehicles to drive themselves
“I mean, I think by 2030, it'll be It's inconceivable to buy a car and not expect it to drive itself.”
Scaringe: Upcoming Rivian R2 will include lidar alongside radar and cameras
“Our approach to this is we have a higher level of capability on our perception stacks. We have better cameras. We have radar. And of course, with R two, we'll have a lidar as well.”
Scaringe: US EV adoption is throttled by a lack of vehicle choices
“So I think my view is the EV adoption in the United States is a reflection of the lack of choice. There's one set of really great choices with Model F and Model Y. I think there needs to be many more.”
Baszucki: Roblox ignores university pedigree as top talent bypasses elite schools
“Well, I think people typically think like I went to Stanford. That's where all the top talent is. We have found that not necessarily to be true. And we have found community college, you know, the small Midwestern engineering school, like, because we're assessi…”
Until aims for reversible whole-body cryopreservation after mastering single organ transplants
“Our long-term goal is reversible whole body cryopreservation for, you know, medical sort of hibernation. But in the near term, what we work on is reversibly cryopreserving human, single human organs to help transplant patients get organs more efficiently.”
Desai: SaaS zero-terminal-value AI bear thesis is overblown
“The, just going on the extreme terminal value being zero for some of the software, they are overblown, ah, and, you know, ah, we'll figure this out together.”
Desai: Fast AI Code Generation Cannot Bypass Enterprise Security and Compliance Barriers
“So yes, wipe coding will allow you to create an app fast. You have a great use case. You have some disruption in mind. That's excellent. But then there is a lot of things that you need from a go-to-market perspective to be able to break in, pass all their chec…”
Desai: MongoDB is the only new database to cross $2B revenue
“There are many attempts have been made by many folks to create next version of database. This and that or next disruptive database. Nobody has passed billion, and definitely nobody has passed two billion, and MongoDB did that.”
Desai: Enterprise feedback on office copilots is poor, unlike coding tools
“A lot of them tried with the office productivity type co-pilots, unclear how much value they got out of it. They're like, okay, this really works with my Excel or, you know, can I really do this thing or create a PowerPoint slide using natural language? The fe…”