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 993 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
Valenzuela: Runway built real-time interactive AI video; deployment depends on unit economics
“I think, I mean, we, we've, we have it. It's more about, like, deploying it and, like, the new economics around it.”
Valenzuela: Runway penetrated major film studios entirely through organic inbound employee demand
“That's how we've got into all the studios. It's not us, like, trying to, like, deeply sell to them. It's, like, they reach out being like, hey, we have this use case, or, like, we want to use it because our team was using it.”
Pedregal: Granola's biggest competitor has always been Apple Notes
“The biggest competitor that Granola had from day one and even today is Apple notes.”
Pedregal: Granola makes most product decisions based on intuition
“So our general our philosophy that's gotten us here, and it may not get us there as we scale, is we make most product and design intuition, sorry, decisions based on Intuition.”
Pedregal: Audio transcription is Granola's largest cost, exceeding LLM inference
“So the most expensive thing about our business is actually transcription. And historically it's been actually transcription and high quality transcription versus LLM inference.”
Pedregal: Granola does not operate with negative gross margins
“We're not at a negative gross margins right now”
Cherny: Claude Code hit its stride with Sonnet 4 and Opus 4
“We started building quad code when it was still Sonnet at 3.5 and it was okay. And then with, you know, 3.6 and 3.7, it was still, it was fine. It was pretty good. But then when Sonnet four and Opus four came out, that's when it really hit its stride and we fe…”
Applied AI startups often lack the security controls enterprise buyers require
“So there are often new startups, less so on the infrastructure side and more so on the applied side, who we would love to buy from, partner with, but they don't have the security protocols and controls in place for us to feel comfortable operating in their sta…”
Rauch: Vercel Fluid Compute only bills active CPU cycles during LLM wait times
“If you're waiting for 15 minutes for OpenAI to respond, and you do a tiny bit of compute, maybe to like transform it into HTML or transform it into an MCP response or, you know, call other systems. We're only going to charge you for the actual CPU cycles that …”
Canva built an internal implementation of the Model Context Protocol
“We have our own model context protocol that we've stood up just in internally, ah, and we're certainly talking to the big vendors about, ah, interactions there that are possible.”
Dohmke: AI was part of Microsoft's 2018 GitHub acquisition memo
“One additional thing was in fact AI and saying, we believe there's a future that where we can train AI models on the graph is I think what we called it of all the source code that is stored on, on GitHub, but also on the relationships between the developers, y…”
Hebbia requires its 100 NYC employees to work in-office 5-6 days
“So we built a team of a hundred amazing incredibly smart folks all five days a week, sometimes six in office in New York city.”
Hebbia is willing to spend $10,000 annually per user on compute
“We'll spend 10,000 dollars in model costs per user a year. We don't care.”
Howard: Answer.ai operates with no offices, roles, or professional managers
“We're not going to have offices and departments. We don't, you know, we don't have any roles. We don't have any professional managers.”
Ramaswamy: Snowflake initially viewed supporting Iceberg as a defensive move
“And Snowflake initially approached this as mostly a defensive move.”
Ramaswamy: Neeva reached a $300M valuation on under $1M revenue
“We raised too much money. Our valuation was three hundred million dollars for a company barely making a million. That just was not okay.”
Ramaswamy originally agreed to stay at Snowflake for six months
“My original agreement with Frank, the Snowflake CEO, was that I would stay with Snowflake for six months.”
Ramaswamy: Neeva failed because it wasn't 10x better than Google search
“The rules that killed Neva, which is, at the end of the day, until AI search came along, simply didn't have a product that was 10 times better than Google search. That's just the reality. That's the reason we couldn't compete.”
Misra: Captions would be cash flow positive without long-term model training R&D
“Today, like, You know, for us, COGS is like the least of our concerns, basically. I mean, to give you an idea, like if we were to remove a lot of the sort of capital investment in you know, model training and things that are long-term value for us, like we wou…”
Ada is shifting from human handoffs to AI delegating tasks to humans
“And so increasingly we're starting to think less in terms of handoff and more in terms of human delegation. And we're very excited about that because what it means is that instead of just simply handing off a conversation to a human agent and having them figur…”
Masad laid off half of Replit to pivot into founder mode
“Early twenty-twenty-four, when I felt like when I did, when we kind of went into founder mode, laid off, you know, half the team leaned up the executive team and the management structure and brought us back to kind of like a small team with a very intense focu…”
Socher: You.com gets more queries through APIs than consumer product
“We definitely get a lot more queries coming in through APIs than we do over like the normal consumer you.com product at this point.”
Olivier Pomel: Y Combinator rejected Datadog over lack of initial product
“We applied to all sorts of incubators, we got to a Y Combinator interview, and then we didn't get into Y Combinator. I have an email from Paul Graham that says You know, a platform is only as good as its first product, and you don't have a first product, blah,…”
Datadog keeps struggling product teams small rather than killing products
“We did not kill products, but we, the way we do it is we start small, and we've waited a long time to scale up.”
Glyman: Ramp's experimental AI agent achieves 60% to 90% task accuracy
“I think right now, and the agent probably works for you know, depending on the nature of the task 60 to 90% of the task of when it's controlled and inside of the RAMP experience. Obviously, when you're dealing with financial products, it's not high enough to g…”
Katz: ClickHouse partnered with Alibaba Cloud on multi-year revenue share agreement
“In China and other parts of Asia, we've partnered with Alibaba, specifically all the cloud for them to redistribute our cloud binary as an enterprise version. And we'll, we have a revenue share agreement in place with them for the next few years.”
Looker's embedded BI drove 33% of revenue with 2 of 750 employees
“So at Looker, when we sold the business, a third of the revenue was embedded. And out of a team of about seven to 750, there were two people working on that product.”
VAST Data achieves stateless compute by exposing data across the network
“Every node can now see the entirety of the data set. Data on low cost flash, metadata on storage class memory, all on the other side of the network, such that they become stateless, And they don't need to talk to each other anymore, and so now you have the mos…”
VAST Data builds its software stack internally instead of using open-source
“Everything is our own or almost everything is our own. We don't use open source nearly as much as other companies do.”
Ezra's AI erases noise to enable faster, reduced-repeat MRI scans
“So what we do at Ezra, or one of the things we do is we are able to do the scan fewer times, which results in a noisier image, but we have trained AIs to essentially learn what noise looks like in an MRI machine and just, like, erase the noise.”
Traynor: Many startups in his angel portfolio build thin AI wrappers
“When I like, you know, I also invest in companies and I talk to companies in my portfolio and a lot of them are kind of building these thin wrapper solutions.”
Traynor: Intercom plans to push customer support automation from 50% to 90%
“Does it like, you know, we have genuinely a sort of full spectrum of how we're going to automate support to get from our current, say, 40, 50% total automation up to like maybe seventies, eighties, nineties, depending on the business.”
Handy: dbt will start exploring visual, GUI-based workflows this year
“And so this year will probably be the year where we start playing around in that space.”
Duderstadt: Nomic's enterprise sales funnel is literally Twitter, Discord, then sales
“Our sales funnel is literally like Twitter to discord to enterprise sale.”
Poolside will launch a direct competitor to GitHub Copilot X in 2024
“But in terms of what we're delivering to the end user, we're starting with a direct competitor to get up Copilot X. And so natural language, multi-turn dialogue system. Where developers give instructions of what they want done while they're in their editor. Th…”
Adobe trained its Firefly AI models exclusively on licensed material
“We wanted to train off of licensed material that we had a license to train off of.”
DeepScribe CEO: AI achieved parity with human scribes, enabling full automation
“A couple months ago for the first time achieved parity with our AI and our expert human scribes, at which point we took them out of the loop. And so now DeepScribe is fully automated”
DeepScribe ended offshore reviewer contracts as clinicians stopped requesting human review
“Clinicians are barely sending the note to the human to review, yet we were keeping that option. We're keeping this like base level contract with our offshore suppliers. So we decided to officially make the move and that's when we started to see complete shift.”
DeepScribe discards LLM output if it lacks classical model validation
“So for a given task, we'll have a classical model typically produce that that same output. And then if the LLM puts in something that the classical model didn't have in its output we'll go ahead and only go with the classical model's output and disregard the L…”
DeepScribe drafts off Nuance's enterprise market education to win health clients
“We really ride off the backs of Nuance, which is the 800 pound gorilla in our space that is trying to do something similar to what we're doing, but they have the distribution advantage. So they've gone and educated all the customers, and we follow on and basic…”
Srinivas: Perplexity rejected CoreWeave due to high pricing
“Yeah, we don't use Corby because we decoded a very high pricing to us and like we decided not to use them.”
Polu: Dust is near repeatable sales but has not reached PMF yet
“I think we are in the, we are on the precipice of repeated sales processes. So that could call PMF. I wouldn't quite qualify it as PMF yet.”
Shah: Hippocratic AI is building its healthcare LLM from scratch
“That's why we're building our own model and we're building it from scratch.”
Shah: Hippocratic AI operates fully in person five days a week
“We're a hundred percent in person. Our company is literally five days a week”
Ada shifts pricing to charge only for automatically resolved support conversations
“We're so big on this measure that not only is it the North Star measure of our entire company, but we're increasingly pricing, enabling our customers to price according to it. So we're actually only going to charge you for a conversation that's automatically r…”