Components of a Frontier AI Slowdown
Slowdowns specify aims, what is to be slowed, and a structure.
A Strange Attractor
Thoughts on AI policy
Subscribe:.rss.atom.json.md.m3u.pls
Live Last read · last published · next check
Latest episodes
AIs as Whistleblowers
AI systems deployed within AI companies should be given whistleblowing channels and be trained to use them.
Alternative Milestones to AGI
Being good at tasks is not the same as being good at learning tasks.
Measuring AI R&D Automation
Automation of AI R&D could be a big deal. We should track its broader consequences.
Transparency into Model Spec Adherence
Developers are increasingly publishing model specs. Developers should also describe their processes for ensuring that models adhere to their specs, and report on the degree of adherence.
How much could content labeling help with AI-enabled advertising scams?
The limited empirical data suggests that content labeling could help somewhat, but would leave potentially tens of billions in losses unaddressed.
How Could Formal Verification Help with Cyber Resilience?
Formal verification could help to secure critical software, but it faces a number of practical barriers. More advanced AI could help lift many, but not all, of them.