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 684 resolved? a statement only gets an assessment when the public record can support or contradict it. opinions and what-ifs never can, and 0 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
Chollet: Base LLMs Score Under 10% on ARC-AGI-1
“So basal alarms were scoring extremely low on V-one, like sub-ten percent, basically. And, I mean, it was true of the original, like, GPT-III actually scoring zero, but that's even true of the latest basal alarms today, you know, as of March.”
Hu: Confluence Labs Saturated the ARC-2 Benchmark at 97% Accuracy
“I actually worked with a company in the winter 26 batch not too long ago called Confluence Lab, which actually ended up saturating the V-two results with 97%, and I think their task cost was a lot more efficient too.”
Humans can learn to control individual cortical neurons within minutes via feedback
“If I put an electrode almost anywhere in your brain and then wake you up in during surgery, And I show you a flashing light that is, that flashes proportionally to how much that neuron is firing, at least almost anywhere in cortex. Within a couple of minutes, …”
Hodak: Prima implant produces normal black-and-white vision in blind patients
“The qualia of Prima is, is normal sight. It's black and white. It's only a, it's a small field of view, but it's vision.”
Hodak: Directly stimulating retinal bipolar cells produces images in the mind
“It was an empirical discovery of our study that if you excite the bipolar cells with an image, you get an image in the mind's eye, because that is clearly the critical processing step in the retina that you wanted to preserve.”
Hodak: Science found optogenetic proteins sensitive to ambient indoor office lighting
“What we were able to do were find optogenetic proteins that are so sensitive that they're sensitive to like indoor office lighting”
Hodak: Science's neural grafts formed biological brain connections in animal models
“But it comes with the potential of growing throughout the brain, forming biological connections all over the place. And I mean, that's what we've seen in, in the animal models. That's not in humans yet”
Poetiq outperformed Gemini 3 Deep Think on ARC-AGI-2 at half the cost
“Yeah, so the interesting thing is that we were half the cost of Gemini Three Deep Think because we were building on top of Gemini Three Pro, which is a much cheaper model. But we still got in the end, a nine percentage point improvement on the official verific…”
Poetiq scored 55% on Humanity's Last Exam, outperforming Claude Opus 4.6
“AI hasn't passed it yet, but we got to 55%, which is almost two percentage points higher than the previous state of the art. Which came out just last week from Anthropic with Claude Opus 4.6. They got 53.1%, and we got 55% on it.”
SemiAnalysis reports Claude Code generates four percent of all public commits
“There were some other stuff from like semi-analysis that four percent of all public commits are made by quad code.”
NASA JPL used Claude Code to navigate the Perseverance Mars rover
“It plotted the course for perseverance, like for like the Mars Rover.”
French-Owen: Claude Code and Codex Use Grep Rather Than Semantic Embeddings
“Like I think cursor takes an approach where they actually do semantic search, where they embed everything and figure out like, Hey, what query is closest to this? If you look at a codex or a cloud code they actually just use like grep.”
French-Owen: Codex Uses Periodic Compaction to Support Long-Running Tasks
“Where it will run compaction, like, periodically after each turn, and so Codex can continue to run for a very long time, and if you look at the percentage in the CLI, you'll see it, like, move up and down as compaction runs.”
Tan: Gamma reached $100M ARR with only 50 employees
“Gamma was interesting to see, like one of the biggest things that they said in their launch that I think is a very good trend is they said they got to a hundred million dollars in ARR with only 50 employees.”
Kaplan: Scaling laws apply to reinforcement learning in AI training
“You can see scaling laws in the reinforcement learning phase of AI training.”
Kaplan: METR found AI task duration doubles roughly every seven months
“And an organization meter studied this very systematically and found yet another scaling trend. They found that If you look at the length of tasks that AI models can do, it's doubling roughly every seven months.”
Kaplan: AI training and inference will see 3x-10x yearly efficiency gains
“I think that over time, as AI becomes more and more widespread, I think that we're going to really drive down the cost of inference and training dramatically from where we are right now. Sort of, three X to 10 X gains algorithmically, and in sort of scaling up…”
Finn: Physical Intelligence adapted its model to an unseen third-party robot
“We're also able to apply that same recipe to robots at other companies. This is a robot that I've actually never seen in person before. They collected data. They sent the data to us. We fine tuned our model on their data. We actually didn't even know exactly h…”
Finn: Full Pre-Training Mixture Boosts Robot Performance Over 20% in Novel Homes
“And we find that these kind of bars on the right, which are excluding data from static robots in labs and environments and so forth reduces performance significantly. So the performance goes down to less than 60% when you exclude that data when evaluated in no…”
Jumper: A two-day open-source hack created the best protein interaction predictor
“The tweet on the left from Yoshitaka Morawaki came out two days after our code was available. We had predicted the structure of individual proteins, but we considered, we were working on building a system that would predict how proteins came together. But, ah,…”
Chollet: Fine-Tuned OpenAI o3 Reached Human-Level Performance on ARC
“So in particular, in December last year, OpenAI previewed its, ah, all three model, and they used a version of it that was, ah, fine-tuned specifically on Arc, and that showed human-level performance on that benchmark versus time.”
Nadella: LinkedIn merged design, engineering, and product into full stack builders
“Like, you know, at LinkedIn, I think they took multiple of these functions the design function, the front end engineer function, the product function, put them all together and said, we're going to have full stack builders.”
Elon Musk: Raw materials are only 1% to 2% of rocket costs
“Actually a rocket, the raw materials of a rocket are only maybe one or two percent of the historical cost of a rocket. So The manufacturing must necessarily be very inefficient.”
Mac Minis are uniquely suited for batch-one local LLM inference
“Mac minis, for example, are a very good fit for some of the LLMs, because it's all, if you're doing batch one inference, this is all super memory bound, so this actually works.”
Thompson: Penicillin was barely used until US WWII industrial scaling
“Like five people have been delivered penicillin in all of human history, 13 years after it was discovered, and the two of them have died. Penicillin has done nothing, and it took a very concerted U.S. Effort to implement penicillin, to grow it in large vats, t…”
Derek Thompson: NIH awards funding to older scientists over younger researchers
“It funds older scientists, much more than younger scientists, even though, I don't have to tell you this, it's often young people who have the most earth-changing ideas, the best paradigm-shifting ideas.”
Thompson: Today's scientists spend 40% of their time on grant paperwork
“A remarkable statistics from the book is that today's scientists now spend 40% of their time filling out paperwork and filing for grants.”
Xu: DoorDash spent millions on pandemic ads promoting competitor platforms
“One was running a national TV campaign where we spent millions of dollars advertising on behalf of the industry, basically saying whether you order on us or any of our peers, just order.”
Xu: Cutting pandemic merchant commissions in half cost DoorDash over $100M
“And the other decision, which we were the only platform to do, we cut our commissions by half, which cost us over a hundred million dollars.”
Field: Designers fiercely rejected Figma's 2016 multiplayer launch
“I remember when we finally launched the GA in October of 2016, I mean, the comments we got were like, if this is the future of design, I'm changing careers. You know, that person commented about it and said, a camel is a horse designed by committee. So it's li…”
Zenefits grew from $20M to $65M before Conrad was ousted as CEO
“And then the next year, our plan was to go from 20 to a hundred. And we ended up going from 20 to 65, and I got fired, ah, basically within a week of the end of the fiscal year.”
Public software firms spend 50% more on sales for less new ARR
“In the public markets, you know, over the last five years, public companies are spending 50% more on, on sales and marketing but they're adding 10% less in new ARR, like with that increase in spend.”
Lieb: Musk said Twitter should make far-left and far-right equally unhappy
“I think Elon explicitly said, when you're using Twitter, you should feel equally bad if you're on the far left or on the far right.”
Hu: Diode demonstrated that OpenAI's o1 can automate PCB system design
“But the thing that they demonstrated now with O-one was actually able to do the system design and component selection, which is crazy. So it would be able to read all the data sheets and select the right components.”
Altman: I wrote OpenAI's early checks before Elon Musk funded scale
“It was definitely helpful that I could just, like, write the early checks for OpenAI, and I think it would have been hard to get somebody else to do that at the very beginning. And then Elon did it a lot at a much higher scale, which I'm very grateful for, and…”
Tan: OpenAI enabled internal model distillation as a developer lock-in strategy
“OpenAI itself has now enabled distillation internal to its own API. So you can use O-one, you can use even GPT-IV or IV-O to distill it down into a much cheaper model that's internal to them, like GPT-IV, IV-O-Mini. And that's sort of their, you know, lock-in …”
Monzo scaled to 500k users with unit economics of -£30 to -£40
“In the early days of Monzo, we were at -30 or -40 pounds per customer per year, and we scaled to half a million customers before we fixed them.”
Caldwell: YC accepts startups with $500K ARR or $1M+ raised
“We have folks that get into YC that have, I don't know, half a million dollars in annual recurring revenue. Also a common one here these days is, oh, I'm too far along because I've already raised some money. We have folks that apply and get accepted to YC that…”
Caldwell: Fast raised $100M and burned it in 10 months
“What did fast do? They raised a hundred million and they burned it in 10 months. They got into a ten million A month burn.”
Caldwell: Brian Armstrong's original Coinbase YC pitch wasn't about buying Bitcoin
“Another thing I want to note is even Brian and his YC application Didn't really believe in his own idea because what he basically applied to YC with was the P to P transfer parts. It wasn't buying Bitcoin and holding it. It was the idea that he could build som…”
Seibel: Stripe deliberately launched as the most expensive payment product
“When the Stripe Finders came to speak at YC, the other detail they added was it was the most expensive product in the market. Like, to make sure people really wanted it, they didn't give it away for free, they didn't price it comparably, they made it more expe…”
Saez-Gil: Less than 2% of $80B carbon market funds forestry
“According to the World Bank last year, over eighty billion dollars went into carbon market initiatives around the world. And yet, when you look at the percentage of that funding that is going to forest conservation and restoration is less than two percent.”
Saez-Gil: New airline regulations demand 160M tons of carbon offsets annually
“All the airlines starting in twenty-twenty-one will have to offset their emissions. This is by regulations, right? Mm-hmm. It's called the Corsia Framework. Yeah. We are talking about a 160 million tons per year that are gonna be demanded by the market.”
Agroforestry can produce higher yields than traditional monoculture agriculture
“It turns out, according to the data I read, that agriculture inside of a forest can actually produce more yield than traditional, you know, monoculture agriculture.”
Betts-Lacroix: Mint scaled by scraping bank websites using voluntarily submitted credentials
“So a friend of mine, Aaron Patzer, created something called Mint and ended up taking a bunch of eyeballs from big banks and other financial companies by hacking, essentially going, doing adversarial interoperability with their websites by scraping, which is so…”
Rossmann: Up to 75% of elite university students have wealthy parents
“Anywhere from 60 to 75% of the students who go to schools like Stanford and MIT come from the top 20% of families by income, and oftentimes you get up to a quarter of the student body that's from the one percent of families by income.”
Rossmann: Make School is the only surviving young dropout startup from its YC batch
“My co-founder and I are the only young founders in our batch whose company is still alive. So when you look at it from that perspective, you look today, what are the companies from my YC batch that still exist and are still successful? There is, there are zero…”