Abstracted abstract:

Frontier models are increasingly trained and deployed as autonomous agents, which significantly increases their potential for risks. One particular safety concern is that AI agents might covertly pursue misaligned goals, hiding their true capabilities and objectives – also known as scheming. We study whether models have the capability to scheme in pursuit of a goal that we provide in-context and instruct the model to strongly follow. We evaluate frontier models on a suite of six agentic evaluations where models are instructed to pursue goals and are placed in environments that incentivize scheming.

I saw this posted here a moment ago and reported it*, and it looks to have been purged. I am reposting it to allow us to sneer at it.

*

Kind of sharing this because the headline is a little sensationalist and makes it sound like MS is hard right (they are, but not like this) and anti-EU.

I mean, they probably are! Especially if it means MS is barred from monopolies and vertical integration.

TIL that television program “The Boys” has an in universe cryptocurrency as a satire of, well, cryptocurrency in general but also specifically that time when DJT was selling NFTs. They occasionally tweet about it.

It has a listing on the “BSCScan” crypto tracker under the name “VTC” so someone might have actually minted it? It might surprise some of you that I have no way of telling the realness of such a thing.

Uncritically sharing this article with naive hope. Is this just PR for a game? Probably. Indies deserve as much free press as possible though.

Followup to part 1, which now has a transcript!

As is tradition, I am posting this link without having listened to it. (too many podcasts)

Found this because an article on Helen Toner popped up in my feed and I wanted to find out more, and boy did I find out more.

Discussion on AI starts at about 17mins. The Bas(ilisk) drop happens at 20:30. Sorry if ads mess up my time stamps. I think this is the second time it’s come up on the show.