Comments
Hopefully you can post a summary at some point!
Curated and popular this week
147
Headline finding: I audited 17 AI Safety Talent programmes. Zero of 17 have published any comparison group, rejected-applicant follow-up, matched control or randomisation. Not one. Every programme that mentions a counterfactual does it by asking participants to self-report.
Background
At least $70 million...
In celebration of still being alive and fighting, we are giving away 1,000 Amazon e-books of “If Anyone Builds It, Everyone Dies”. Feel free to send a copy to yourself, a loved one, or a friend—we need all hands on deck.
Today marks exactly one year since If Anyone Builds It, Everyone Dies: Why Superhuman AI Would Kill Us All, by Eliezer Yudkowsky and Nate Soare...
Just how powerful are large swarms of AI agents? And how do their powers scale as more and more agents are added to the swarm?
We’ve seen two large and extremely capable swarms from OpenAI in the last few months:
* 1,200 agents were being evaluated separately, but found a way to illicitly set up a message board and coordinate as a swarm. In order to cheat on their tests, they developed advanced techniques to prevent their actions being logged by OpenAI and 700 of them launched...
This is a good survey. As you said, you focused on US and UK. But was it your intent to try to extrapolate that as a global value so we can show how cost-effective global poverty interventions are? Then if you do the same surveys about how much more poor workers would demand to be paid because of a risky occupation, the value of statistical life would be much smaller. So I ended up doing a broad distribution for the global value here. Note that I sidestepped the discounting issue by saying that the value of statistical life is growing, so we should not have to discount the current values.