| Not yet due |
Prediction |
11 Nov 2024 |
Amodei said that probably by 2027 there will be ambitions to build hundred-billion-dollar compute clusters, and that he thinks this will actually happen.
Prediction · 11 Nov 2024 |
AI compute, hardware and data |
| Not checkable |
Not checkable |
11 Nov 2024 |
I think releasing the model while the capabilities are still limited is very helpful in terms of doing that.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
we also need to address the risks because those risks are in our way. They’re landmines on the way from here to there, and we have to…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
We've seen lots of examples of demagoguery in our life from humans, and there's a concern that models could do that as well.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
AI is not God
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
It’s about shaping the incentives to point upward instead of shaping the incentives to point downward.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
In many ways, the manner and personality of these models is more an art than it is a science.
Not checkable · 11 Nov 2024 |
|
| Not researched |
Factual |
11 Nov 2024 |
Performance on a programming benchmark has gone from succeeding about 3% of the time to about 50% of the time.
Factual · 11 Nov 2024 |
AI capabilities |
| Not checkable |
Not checkable |
11 Nov 2024 |
probably going forward, we might release new ones multiple times a year because it's hard to get these policies right technically…
Not checkable · 11 Nov 2024 |
|
| Not researched |
Factual |
11 Nov 2024 |
Anthropic is not currently building its own coding IDEs but powers companies such as Cursor and Kognition that build them on its API.
Factual · 11 Nov 2024 |
Their own work or organisation |
| Not checkable |
Not checkable |
11 Nov 2024 |
I think that's like a present day analog of future control problems in AI systems that we can start to study today.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I think containing bad models is a much worse solution than having good models.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I just saw these as independent dials that you could turn.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
That in itself I think can overcome almost every other disadvantage.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
we have a theory of change called Race to the Top. Race to the Top is about trying to push the other players to do the right thing by…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I’m not at all confident.
Not checkable · 11 Nov 2024 |
|
| Not researched |
Factual |
11 Nov 2024 |
Anthropic's interpretability research had no commercial application for three or four years and still has none.
Factual · 11 Nov 2024 |
Their own work or organisation |
| Not checkable |
Not checkable |
11 Nov 2024 |
It seems like the process is what matters and how it shows who you are as a person along the way and how you relate to other people and the…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
So I actually do believe that you can gain benchmarks, but I think if we get to 100% on that benchmark in a way that isn’t over-trained or…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
But the system as a whole, I think makes some obvious trade-offs that are very far from maximizing human welfare.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
In particular, linear scaling up of bigger networks, bigger training times and more and more data.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I could point you to a hundred people here who are better at that than I am.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
generally, you want to preserve mechanistic interpretability as a verification set or test set that's separate from the training process of…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
If you have a strong pre-trained model, I feel like you’re halfway to anywhere in terms of the intelligence space.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I think the time scale is just too long and I can see it.
Not checkable · 11 Nov 2024 |
|
| Resolved true |
Factual |
11 Nov 2024 |
On some computer-use benchmarks, previous models succeeded about 6% of the time and Anthropic's new model succeeds 14 or 22% of the time.
Factual · 11 Nov 2024 |
AI capabilities |
| Not researched |
Factual |
11 Nov 2024 |
OpenAI built GPT-3, the model that was commercialized, and Dario Amodei was involved in its commercialization.
Factual · 11 Nov 2024 |
Their own work or organisation |
| Not yet due |
Prediction |
11 Nov 2024 |
Powerful AI's transformative deployment in the world will take more likely five or 10 years than 50 or 100 years.
Prediction · 11 Nov 2024 |
AI timelines and forecasting |
| Not checkable |
Not checkable |
11 Nov 2024 |
I think that’s a pretty useful direction.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
these systems are actually quite unpredictable. They're actually quite hard to steer and control.
Not checkable · 11 Nov 2024 |
|
| Not researched |
Factual |
11 Nov 2024 |
Anthropic was the only AI company that felt positively about California's SB 1047 in a very detailed way, while some of the big companies, such as Google, OpenAI, Meta and Microsoft, were pretty staunchly against it.
Factual · 11 Nov 2024 |
AI policy and governance |
| Not researched |
Factual |
11 Nov 2024 |
Asked about the limits and expense of compute, Amodei guessed that most frontier model companies are operating at roughly a 1 billion scale, plus or minus a factor of three.
Factual · 11 Nov 2024 |
Economics and markets |
| Not researched |
Factual |
11 Nov 2024 |
Claude's computer use feature adds that input images can be computer screenshots, with the model trained to output screen click locations or key presses.
Factual · 11 Nov 2024 |
AI capabilities |
| Not checkable |
Not checkable |
11 Nov 2024 |
The perfect organization doesn’t exist.
Not checkable · 11 Nov 2024 |
|
| Unresolvable |
Prediction |
11 Nov 2024 |
For much longer than expected, the small parts of jobs humans still do will expand to fill their whole jobs as AI automates the rest.
Prediction · 11 Nov 2024 |
AI, the economy and work |
| Resolved true |
Factual |
11 Nov 2024 |
Anthropic's updated Claude Sonnet 3.5 scores about 50% on SWE-bench.
Factual · 11 Nov 2024 |
AI capabilities |
| Not checkable |
Not checkable |
11 Nov 2024 |
I think we were in a good position in terms of naming when we had Haiku, Sonnet and Opus.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
once I got to biology sections, I wasn’t an expert.
Not checkable · 11 Nov 2024 |
|
| Not researched |
Factual |
11 Nov 2024 |
Sonnet 3.5 has roughly the same cost and speed as Sonnet 3 but is smarter than the original Opus 3, especially for code.
Factual · 11 Nov 2024 |
AI capabilities |
| Not checkable |
Not checkable |
11 Nov 2024 |
I am optimistic about meaning. I worry about economics and the concentration of power. That’s actually what I worry about more, the abuse…
Not checkable · 11 Nov 2024 |
|
| Unresolvable |
Prediction |
11 Nov 2024 |
Dario Amodei said that as a CEO he does not get to program much, and that he felt that if he went back to it six months later, it would be completely unrecognizable to him.
Prediction · 11 Nov 2024 |
AI capabilities |
| Not checkable |
Not checkable |
11 Nov 2024 |
So we’re on that S-curve where it’s going to start slowing down soon because you can only get to 100%.
Not checkable · 11 Nov 2024 |
|
| Not researched |
Factual |
11 Nov 2024 |
Anthropic has had to change its Responsible Scaling Policy framework, releasing a new version a few weeks before November 2024.
Factual · 11 Nov 2024 |
Their own work or organisation |
| Not checkable |
Not checkable |
11 Nov 2024 |
Anthropic will be successful.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
We are rapidly running out of truly convincing blockers, truly compelling reasons why this will not happen in the next few years.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
That’s less important than, again, that we align the incentives of the industry.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I don’t think you can scale up humans enough to get high quality.
Not checkable · 11 Nov 2024 |
|
| Not yet due |
Prediction |
11 Nov 2024 |
Over the next year or two, even without better models, there would be enormous opportunity to boost productivity by catching mistakes and doing grunt work with current AI models.
Prediction · 11 Nov 2024 |
AI, the economy and work |
| Not checkable |
Not checkable |
11 Nov 2024 |
this idea that within large organizations and systems, there end up being a few people or a few new ideas who cause things to go in a…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
If we’re going to have a sandbox, it would need to be a mathematically provable.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I will say, while in theory there’s nothing you could do there that you couldn’t have done through just giving the model the API to drive…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
But the thing that I think I did have that was different was that I was just willing to look at something with new eyes.
Not checkable · 11 Nov 2024 |
|
| Mostly true |
Factual |
11 Nov 2024 |
Anthropic wants to evaluate every new model to see whether it is getting close to dangerous (CBRN) capabilities, and does not think current models yet pose these risks seriously.
Factual · 11 Nov 2024 |
Biosecurity |
| Not checkable |
Not checkable |
11 Nov 2024 |
It’s more like it kind of opens the aperture for the model to use and apply its existing abilities.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
The talent mass is greater in the group of a thousand people.
Not checkable · 11 Nov 2024 |
|
| Unresolvable |
Prediction |
11 Nov 2024 |
Scaling up models will enable AI to perform very wide-ranging cognitive tasks.
Prediction · 11 Nov 2024 |
AI capabilities |
| Not checkable |
Not checkable |
11 Nov 2024 |
If you extrapolate the curves that we’ve had so far, right?
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I think in terms of drug development, my view is that we’re too slow and we’re too conservative.
Not checkable · 11 Nov 2024 |
|
| Not yet due |
Prediction |
11 Nov 2024 |
Eyeballing the rate at which AI capabilities are increasing makes Amodei think "we'll get there" by 2026 or 2027, though he says there are still worlds where it does not happen in 100 years.
Prediction · 11 Nov 2024 |
AI timelines and forecasting |
| Not checkable |
Not checkable |
11 Nov 2024 |
I think drawing that line as finely as possible, approaching perfectly, is still a challenge and we're getting better at it every day
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I don’t think you can trust these companies to adhere to these voluntary plans on their own.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
There’s no reason why they should be designed for us to understand them, right?
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
our view has been let 1,000 flowers bloom.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
basically it’s going to need to follow basic human laws.
Not checkable · 11 Nov 2024 |
|
| Mostly true |
Factual |
11 Nov 2024 |
Amodei first joined the AI field in late 2014, working at Baidu with Andrew Ng.
Factual · 11 Nov 2024 |
Their own work or organisation |
| Not checkable |
Not checkable |
11 Nov 2024 |
I think RL has the potential to make models smarter, to make them reason better, to make them operate better, to make them develop new…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
Somehow these interventions on the model, where you kind of adjust its behavior, somehow emotionally made it seem more human than any other…
Not checkable · 11 Nov 2024 |
|
| Not yet due |
Prediction |
11 Nov 2024 |
Within two to three years, whether or not super powerful AI exists, compute clusters will be large enough to deploy millions of AI model instances.
Prediction · 11 Nov 2024 |
AI compute, hardware and data |
| Not checkable |
Not checkable |
11 Nov 2024 |
It has a strong personality. It has these kind of obsessive interests.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
Actually, it's just very hard to control the behavior of the model, to steer the behavior of the model in all circumstances at once.
Not checkable · 11 Nov 2024 |
|
| Unresolvable |
Prediction |
11 Nov 2024 |
AI will disrupt agriculture, though more slowly than skills closer to AI builders.
Prediction · 11 Nov 2024 |
AI, the economy and work |
| Not checkable |
Not checkable |
11 Nov 2024 |
I’m caricaturing this a little bit, but I think that’s one extreme.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I think we’ve been pretty good in general at RL, perhaps the best, although I don’t know, I don’t see what goes on inside other companies.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
You got to give this thing boundaries and guardrails.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I don’t think it makes the model smarter.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
RLHF is good at doing what humans want the model to do, or at least to state it more precisely doing what humans who look at the model for…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
Individual companies can help to start it, can help to accelerate it.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
ASL-1 is for systems that manifestly don't pose any risk of autonomy or misuse. So for example, a chess playing bot, Deep Blue would be…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
probably I said a bunch of things that were embarrassing or wrong.
Not checkable · 11 Nov 2024 |
|
| Unresolvable |
Prediction |
11 Nov 2024 |
Running out of training data could become a limit on AI scaling, though Amodei would bet against it.
Prediction · 11 Nov 2024 |
AI compute, hardware and data |
| Not checkable |
Not checkable |
11 Nov 2024 |
You have very little ability to see what’s going on and even less ability to change it, right?
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
They’re complex enough that just iterating, just running the experiment is going to beat any modeling, no matter how smart the system doing…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I don't want to commit to any naming scheme because if I say here, "We're going to have Claude 4 next year," and then we decide that we…
Not checkable · 11 Nov 2024 |
|
| Not researched |
Factual |
11 Nov 2024 |
With each new model, every few months, Anthropic measures that its models are getting better at concerning tasks as well as at valuable, economically useful tasks.
Factual · 11 Nov 2024 |
AI safety and alignment |
| Not checkable |
Not checkable |
11 Nov 2024 |
It's a multidimensional problem.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
And the whole reason we’re trying to prevent these risks is not because we’re afraid of technology, not because we want to slow it down.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
If instead you create a race to the top where people are competing to engage in good practices, then at the end of the day, it doesn’t…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I really think we need to do something in 2025.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I think all these questions of drama are profoundly uninteresting and the thing that matters is the ecosystem that we all operate in and…
Not checkable · 11 Nov 2024 |
|
| Unresolvable |
Prediction |
11 Nov 2024 |
Dario Amodei said he would suspect that if AI models reach 90-95% on the software-engineering benchmark, that will represent the ability to autonomously do a significant fraction of software engineering tasks.
Prediction · 11 Nov 2024 |
AI capabilities |
| Not checkable |
Not checkable |
11 Nov 2024 |
And if by AGI you mean AI is getting better and better, and gradually it’s going to do more and more of what humans do until it’s going to…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
But there’s some cultural trade craft of how we think about the design process that I think is more important than any particular gizmo…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
it creates this negative externality.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
So I think RLHF unhobbles the models in some ways.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
There's actually a huge distribution shift between the stuff that people complain loudly about on social media and what actually…
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I could imagine if we handle things badly in an AI world, we could set things up where people don’t have any long-term source of meaning
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
They’re empirical regularities.
Not checkable · 11 Nov 2024 |
|
| Not checkable |
Not checkable |
11 Nov 2024 |
I think once we get to ASL-4, we start to have worries about the models being smart enough that they might sandbag tests, they might not…
Not checkable · 11 Nov 2024 |
|
| Resolved true |
Factual |
11 Nov 2024 |
Anthropic has almost a thousand employees, and its risk policy is meant to make sure they understand that taking these risks seriously is one of the company's highest priorities, if not the highest.
Factual · 11 Nov 2024 |
Their own work or organisation |
| Not checkable |
Not checkable |
11 Nov 2024 |
you need tests to tell you when the risk is getting close. You need an early warning system.
Not checkable · 11 Nov 2024 |
|