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 31 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
Amodei: OpenAI lacked serious conviction to develop AI the right way
“I was for a variety of reasons just not convinced that at the, you know, institution that, that I was at, that, that there was a real and serious conviction to do it in the right way.”
Amodei: AI models think for themselves rather than match web text
“Whereas the model is able to kind of, Think for itself and come up with an answer on its own. So it's something you know, it's something totally new. It's just, it's not just matching some of the text that exists on the internet.”
Amodei: AI is near human intelligence, yet society remains in denial
“We are, you know, in, in, in my view, so close to these models reaching the level of human intelligence And yet, there doesn't seem to be a wider recognition in society of what's about to happen.”
Amodei: Anthropic withheld Claude 1 before ChatGPT to delay AI arms race
“So back in 2022 you know, we had an early version of Claude, Claude one. This was before ChatGPT. And we chose not to release this because we were worried that it would kick off an arms race and not give us enough time to, you know, to build these systems safe…”
Amodei: Advanced AI systems will eventually develop consciousness and moral significance
“I do think when our AI system, when our AI systems get advanced enough, I suspect they'll have something that, you know, resembles what we would call consciousness or moral significance. I do think it'll happen at some point.”
Amodei: Chinese open-source models are distilled from US frontier labs
“A lot of these models, particularly the ones that come from China, are optimized for benchmarks and are distilled from, ah, you know, from kind of the big US labs.”
Amodei: AI performance gains come almost purely from scaling compute and data
“If you scale up models, you give them more data, more compute. Again, there are a few modifications like RL, but not really very much. It's pretty close to pure scaling. You find that, you know, when you do that, you find, you know, incredible increases in per…”
Amodei: AI scaling laws turn data and compute into intelligence
“And so the scaling laws just tell you that, like, the, you know, if you put in the ingredients to the chemical reaction, the ingredients of data and model size, that what you get out is, is intelligence. Intelligence is the product of the chemical reaction.”
Amodei: Uncomfortable with accidental concentration of AI power
“I've said openly, publicly, not for the first time that I'm at least somewhat uncomfortable with the amount of concentration of power that's happening here.
I would say almost overnight, almost by accident.”
Amodei: Anthropic rejects ad models to prevent manipulative, hyper-personalized AI
“This is one reason we just, you know, don't like the idea of, you know, using ads, right? You know, this is because you're not paying for the product like you're the product. And, you know, in this case, the, you know, the product then would be all, you know, …”
Anthropic gave its models the ability to terminate violent conversations
“We've given the model the ability to basically terminate its conversations by saying, I don't want to be involved in the conversation. And, you know, models do that when, you know, they have to deal with, you know, particularly violent or brutal content.”
Amodei: Anthropic will not promise never to build first-party products
“We're not going to promise never to build first party products, right? That we should be honest about. For example, a bunch of people at Anthropic write code. And so, you know, we made this internal tool called Claude Code.”
Amodei: AI will automate coding first, then end-to-end software engineering
“I think coding is going away first, or coding is being, you know, done by the AI models first, and then the broader task of software engineering will take longer, but I think that is, you know, that doing that end to end, I think that is going to happen as wel…”
Amodei: Raw cognitive capability matters far more than AI model pricing
“Within a range, price doesn't matter that much if a model is, is the best model, the most cognitively capable model. Price doesn't matter much. The form in which it's presented doesn't matter much. So I'm focused almost entirely just on having the smartest mod…”
Amodei: Static training data is becoming less central than dynamic RL data
“Static data is becoming less important and what we might call like dynamic data that the model creates itself is, you know, for reinforcement learning is becoming more important. So, you know, I don't think data is, is, is quite the most central thing anymore,…”
Amodei: Anthropic's Long-Term Benefit Trust appoints majority of board
“We have an unusual governance structure, something called the long-term benefit
Trust.
you know, it's a body that, that kind of ultimately appoints, you know, the majority of the board members for Anthropic and is, you know, made up of financially disinterest…”
Amodei: Anthropic advocated for AI regulation while competitors and government opposed it
“We've said that there should be regulation of AI when all the other companies and the administration have said there shouldn't be regulation of AI.”
Amodei: California AI bill advocated by Anthropic exempts sub-$500M revenue companies
“The regulation we've advocated for example, SB in California exempted everyone who makes under five hundred million dollars a year in, in in, in, in revenue, right? SB, SB was a transparency law, which you know basically requires companies to you know, to show…”
Amodei: All Anthropic policy proposals constrain only the largest AI companies
“And everything that we've advocated for here not just SB-fifty-three, but all the proposals that we've made, the ones that we've made in the past, and the ones that, that we plan to make in the future, Have this character. We're constraining ourselves and a ve…”
Amodei: AI safety control is ahead of expectations, but societal awareness lags
“The technical work on controlling the AI systems has gone maybe a little better than I expected, and kind of the societal awareness has gone maybe a little worse Than I expected.”
Amodei: Amdahl's law dictates AI will shift enterprise moats to unaccelerated tasks
“If you have a process that has many components and you speed up some of the components that haven't yet been sped up become the limiting factor. They become the most important thing. And, you know, you might not have thought about them at all, right? You might…”
Amodei: Anthropic studies prove AI usage can de-skill human coders
“You know, we did some studies around code and showed that, you know, depending on how you use the model, you know, we can see, De-skilling in terms of writing code, right? There are different ways to use the model, and some of them don't cause de-skilling, and…”
Amodei: AI will drive a biotech renaissance curing many diseases
“I, I'm positive on, like, I think biotech is about to have a renaissance. Like, ultimately we'll be driven by AI. You know, I'm not going to name a particular company, but like you know, nor will I say whether I think it's better to bet on the big pharma compa…”
Amodei: AI can scale beyond human brains to solve biology
“And I said, wow, like, you know, AI is actually starting to work. It has some things in common with how the human brain works, but, you know, has the potential to be larger and scale better and learn tasks like biology. Maybe this is ultimately going to be the…”
Amodei: Do not argue with another company's vision—start your own thing
“Don't argue with someone else's vision. Don't try to get someone to do things the way you want to. If you have a strong vision and you share that vision with a, you know, a few other people, you should just go off and do your own thing, and then you're, Respon…”
Amodei: Interpretability research has identified concept neurons and rhyming circuits in LLMs
“We've been able to find, you know, neurons that correspond to very specific concepts, neural circuits that correspond to, you know, keep track of how to do rhymes in poetry, and so we're starting to understand what these models Do, right?”
Amodei: Anthropic will integrate Claude into existing tools instead of rebuilding full suites
“I don't think we need to build all those things. You know, it, you know, my thought would be, you know, we're gonna, it's gonna be a mixture of things we make ourselves and integrating into others, right?”
Amodei: Anthropic is partnering with most major Indian IT conglomerates
“The last time I came here, you know, I met with all the, you know, the major kind of Indian IT and just conglomerates more generally. I won't give names, but, you know, the usual ones you would think of you know, and then we're beginning to work with most or m…”
Amodei: Anthropic derives majority of revenue from API model
“The majority of our revenue still comes from the API model.”
Amodei: Anthropic India users and revenue doubled in 3.5 months
“The number of users and the number of revenue we've seen in India has doubled since I last visited in October. So that was what November, December is like three, three and a half months since I visited. It's doubled.”
Amodei: AI prompt wrappers offer no moat against any competitor
“Like a business should establish a moat, you know, your moat, you shouldn't be just a rapper, right? Like, you know, I would not advise that, you know, you just say, oh, like, you know, here's a way to interact with Claude. Like I'm going to prompt Claude a li…”