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 45 resolved? a statement only gets an assessment when the public record can support or contradict it. opinions and what-ifs never can, and 3 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
Weinberg: AI won't replace junior lawyers, law firms will do more work
“I don't think so. I think we'll just get more work.”
Weinberg: VCs give bad hiring advice because executives manage them up
“The thing that I think they've been wrong about is who to hire. And I think sometimes the problem that VCs have is they're managed up. Right? Like they don't actually see inside a lot of these businesses. They see the board meetings. Right? And so sometimes th…”
Weinberg: Most enterprise customers do not know Sequoia or Andreessen Horowitz
“The vast majority of our customers don't know who Sequoia, A-sixteen, Z, or any of those people even are, right?”
Weinberg: EQT brand gives legal tech startups more credibility than Silicon Valley VCs
“A vast majority of VCs give you that brand, and what's actually interesting for us is, like, someone like EQT actually gives you that more than Silicon Valley because they're private equity, and, you know, a lot more lawyers know who that is, et cetera, right?”
Weinberg: Product iteration speed is the main threat to application AI
“I think just moving fast enough on product, like that, that is always, I think, the biggest existential threat for all the application layer companies. And it's not necessarily that, you know, Anthropic or OpenAI are, you know, tomorrow gonna put 50% of their …”
Weinberg: Consumer AI is plateauing because reasoning is already sufficient
“So I think that we're seeing a plateau in performance for consumer use cases. And the reason why I think this is like a misnomer or something that people actually shouldn't pay attention to is we don't need them to be better for consumer use cases. Like, I fee…”
Weinberg: Code generation and enterprise AI will not hit a performance plateau
“On the enterprise, I think things are going to keep going, and especially Cogen. I think we are not going to see a plateau in Cogen. I think that is going to get much better really, really fast.”
Weinberg: OpenAI and Anthropic revenues would explode even without new models
“I think if both companies literally just stopped shipping things, their revenues would still explode because there are going to be so many companies building on top of their models There are going to be so many different parts of the economy that adopt these t…”
Weinberg: Massive enterprise AI productivity gains are 3-5 years away
“I think like three to five years until we see like massive, massive productivity gains in enterprise.”
Weinberg: B2B SaaS value will explode by aligning with customer ROI
“I think the value of B to B SaaS is about to become astronomical. And if you can figure out how to align your product to that ROI, it's not a hostage. It's, you're just completely aligned with your customer.”
Weinberg: In chaotic markets like AI, VCs rely on resumes over judgment
“When there are situations of chaos and folks don't know what's going on, the safest thing is to go look at other social signals to make decisions instead of using your own gut. And so I think that that's happening a lot in AI, and hopefully, you know, as the m…”
Weinberg: There are only a few hundred truly great AI researchers globally
“Hundreds, and that's it.”
Weinberg: VCs Cannot Identify Top AI Researchers from Resumes Alone
“I don't think you can. That, that I think is actually a pretty big disconnect. The researcher community knows. So like, this is a huge disconnect between VCs, I think, and the researcher community.”
Weinberg: Abandon standard negotiation tactics when you know a deal's core value
“Know when to not negotiate. This, I think, is actually really, really important, and the best deal makers I know are very good at this, which is there are certain deals where you want one thing from the deal and nothing else matters. This only works when you u…”
Weinberg: Founders should not negotiate compensation with best-in-class job candidates
“If you want to hire somebody, hire them, whatever they want to be hired, and put them in the position that they want. If they're best in class, if you can't tell if they're best in class, it's a separate problem, but don't go back and forth. It doesn't matter.”
Weinberg: VCs accepting reduced ownership stakes don't believe the startup will win
“Oh, then you don't believe in that company. That's not going to be a legendary category defining company. It won't be.”
Weinberg: Harvey's annual ARR target significantly exceeds 2.5x growth
“Our goal is much, much higher than that and I think we can do better than that this year.”
Weinberg: Delaying Executive Hires Created Unnecessary Competitors for Harvey
“I took too long to hire senior execs in some instances, and it caused us problems. It created competitors when there shouldn't have been competitors, things like that.”
Weinberg: Candidates joining for elite VCs rarely care about startup mission
“The reality is that might be the wrong person to hire in the first place, because the people that think or go to a company because of the investors usually don't care that much about the mission of the company.”
Weinberg says Harvey's query routing to Anthropic Opus 4.5 rose significantly
“Not entirely, but there were definitely, so we route basically based off of the use case, we route to the best combination of models and definitely our traffic went up to Opus 4.5 significantly.”
Weinberg: OpenAI's investment agreement does not mandate exclusive model usage
“There's nothing in like our agreement or our relationship or anything like that, that you have to use OpenAI models.”
Weinberg: Number two companies gain distribution by attaching to market leaders
“I think that if you are number two in the market, one of the things that's, that can get you a lot of attention is just attaching yourself to number one in any way, shape, or form, right? You kind of get free press from doing that type of thing. It's a good wa…”
Weinberg: Harvey will transition product areas to consumption-based pricing
“There are a lot of areas where we are going to start moving to consumption based pricing. And I think for, at least for us, for our customer base, that would be completely fine with them.”
Weinberg: In-House AI Talent Will Become Critical Again as Startups Scale
“And as these companies scale and they have to create more differentiated products, I think having AI talent is going to matter again.”
Weinberg: Early Harvey AI model produced perfect legal answers for 86% of questions
“And of 86 out of a hundred questions, three out of three said, this is a perfect answer.”
Weinberg clears his Slack inbox every 15 minutes across all channels
“One of the ones I have that I think was a good habit in the beginning of the company is now, and now it's starting to get bad, is I zero out Slack, like, every, like, 15 minutes, and so I'm in, like, almost every single Slack channel, and I read every single t…”
Weinberg: Fast-growing startups cycle through product-market fit and company-market fit
“I see, like, stage one as product market fit, right? Stage two is, like, company market fit. In other words, like, have you created the structures of your company that are the same as traditional, you know, B to B SaaS or whatever, consumer SaaS, whatever you'…”
Weinberg says Anthropic reached roughly $3 billion in revenue entering 2025
“I'm pretty sure Anthropix at like three billion in revenue right now.”
Weinberg: AI market winners and losers will be decided within two years
“Like I think the winners and losers are going to be decided in the next couple of years, right?”
Weinberg: Harvey targets only one or two investors when fundraising
“What I actually do is there's a couple investors that I've become decently close with, and when it's time to fundraise, Usually we get preempted or I say, Hey, I think it's like time to fundraise. And I reach out to literally like one or two people.”
Weinberg: Starting fundraising 6 months early requires less work than standard processes
“I think the main thing about fundraising is you should always think of it as start at six months ahead of time, and you will do much less work than if you actually go out and do the process, right?”
Weinberg: OpenAI's consumer brand power outside tech is unbelievably dominant
“The main thing with OpenAI is they have so much consumer brand. It is so powerful, and especially outside of X and outside of kind of like our worlds. It is so unbelievably powerful, and I don't come from the tech world. And so all of my friends, all of my rel…”
Weinberg: Non-legal corporate teams adopt Harvey without dedicated features
“Starting to happen is those departments are adopting Harvey, even though we haven't built features specifically for those departments.”
Harvey deployed a 4,000-person enterprise rollout with just four employees
“I mean, we were like, when we signed Anno Sherman, which was our first customer, we had four people. So we did a 4000 person enterprise grade rollout with four people.”
Weinberg: 90% of engineers at AI application startups are front-end hires
“When I look at a lot of AI application layer companies, if you go through their LinkedIn's and you look at the engineers that they're hiring, it's like a 90% front end engineers, which is interesting to me.”
Weinberg: 60% of Harvey's revenue comes from law firms, 40% in-house
“I think it's around 40% of our revenue is in-house corporate, and 60% is law firm. Something like that.”
Weinberg: AI Researchers Leave Big Labs When Strategic Direction Changes
“Researchers really genuinely do care about pushing research forward, and I think what ends up happening sometimes is the leaders at some of these companies kind of change the direction of the company, and the researchers are no longer interested in that direct…”
Weinberg: Harvey raised its seed round exclusively from OpenAI
“When we did our seed, and we didn't go to anybody else. It was just OpenAI.”
Weinberg reveals Harvey's seed round pre-money valuation was around $4 million
“The actual terms, I don't remember what the post actually was, but the pre-investment was like, it was like four million or something like that.”
Harvey users utilizing four or more products reach 74% DAU/MAU
“Our DAU over MAU for people that use four plus product lines is 74%.”
Weinberg reveals OpenAI, Sarah Guo, and Elad Gil as Harvey's first investors
“So our first investor was OpenAI and then our first two angel investors were Sarah Guo and Elad Gil.”
Weinberg: AI leaders must benchmark against broader market pull, not just competitors
“One of the biggest jobs that you have is to make sure that your team doesn't feel like they've already won, right? Because the reality is the market pull is massive, and so sometimes your success isn't just your execution, it's the market pull, right? And so y…”