Joe summarizes his new report on "scheming AIs" - advanced AI systems that fake alignment during training in order to gain power later. He explores different types of scheming (i.e. distinguishing "alignment faking" from "powerseeking"), and asks what the prerequisites for scheming are and by which paths they might arise.
I've recently written about how I've updated against seeing substantially faster than trend AI progress due to quickly massively scaling up RL on agentic software engineering. One response I've heard is something like:
RL scale-ups so far have used very crappy environments due to difficulty quickly sourcing enough decent (or even high quality) environments. Thus, once AI companies manage to get their hands on actually good RL environments (which could happen pretty quickly), performance will increase a bunch.
Another way to put this response is that AI companies haven't actually done a good job scaling up RL—they've scaled up the compute, but with low quality data—and once they actually do the RL scale up for real this time, there will be a big jump in AI capabilities (which yields...
I agree with this directionally, and generally tend to believe that certain kinds of straight lines will continue to be straight. But also iirc you rely a lot on the METR trend line, which is (a) noisy, (b) over a short timescale (~2 years[1]), and (c) not a great measure of what people care about[2][3]. I don't think you should strongly expect that particular line to continue to be straight over a long time period.
(Though I tend to think that it's more likely to slow down by 2027 than to speed up, relative to what you'd predict by extrapolating the 2024-2...
1While I do spend some time discussing AGI timelines (and I've written some posts about it recently), I don't think moderate quantitative differences in AGI timelines matter that much for deciding what to do[1]. For instance, having a 15-year median rather than a 6-year median doesn't make that big of a difference. That said, I do think that moderate differences in the chance of very short timelines (i.e., less than 3 years) matter more: going from a 20% chance to a 50% chance of full AI R&D automation withi...
41
1Here's a relatively important question regarding transparency requirements for AI companies: At which points in time should AI companies be required to disclose information? (While I focus on transparency, this question is also applicable to other safety-relevant requirements, and is applicable to norms around voluntary actions rather than requirements.)
A natural option would be to attach transparency requirements to the existing processes of pre-deployment testing and releasing a model card when a new model is released. As in, companies would be required to include the relevant information whenever they release a new model (likely in the model card). This is convenient because pre-deployment testing and model cards are already established norms in the AI industry, which makes it easier to attach something new to these existing processes rather...
One possible important way to address parts of this is by moving from only thinking about model audits and model cards, towards organizational audits. That is, the organization should have policies about when to test and what and when to disclose test results; an organizational safety audit would decide if those policies are appropriate, sufficiently transparent, and sufficient given the risks - and also check to ensure the policies are being followed.
Note that Anthropic has done something like this, albeit weaker, by undergoing an ISO management sys...
2(Last revised: June 2025. See changelog at the bottom.)
Part of the “Intro to brain-like-AGI safety” post series.
(Much of this post had been originally published here.)
The previous post presented a big picture of how I think motivation works in the human brain, but it was a bit abstract. In this post, I will walk through an example. To summarize, the steps will be:
Thanks! I’m not 100% sure what you’re getting at, here are some possible comparisons:
(Cross-posted from X, intended for a general audience.)
There’s a funny thing where economics education paradoxically makes people DUMBER at thinking about future AI. Econ textbooks teach concepts & frames that are great for most things, but counterproductive for thinking about AGI. Here are 4 examples. Longpost:
THE FIRST PIECE of Econ anti-pedagogy is hiding in the words “labor” & “capital”. These words conflate a superficial difference (flesh-and-blood human vs not) with a bundle of unspoken assumptions and intuitions, which will all get broken by Artificial General Intelligence (AGI).
By “AGI” I mean here “a bundle of chips, algorithms, electricity, and/or teleoperated robots that can autonomously do the kinds of stuff that ambitious human adults can do—founding and running new companies, R&D, learning new skills, using arbitrary teleoperated robots after...
This is the idea that at some point in scaling up an organization you could lose efficiency due to needing more/better management, more communication (meetings) needed and longer communication processes, "bloat" in general. I'm not claiming it’s likely to happen with AI, just another possible reason for increasing marginal cost with scale.
Hmm, that would apply to an individual firm but not to a product category, right? If Firm 1 is producing so much [AGI component X] that they pile up bureaucracy and inefficiency, then Firms 2, 3, 4, and 5 will start produ...