"congratulations" danielle, you've successfully gotten me to go around anthropic slack going "nooo you don't understand, Death is actually a VERY GOOD card"

from a research paper 🥲

Back in 2016, I asked coworkers aiming to "build AGI" what they thought would happen if they succeeded. Some said ~"lol idk". Dario said "here's some long google docs I wrote". He does much more "writing-to-think" than he publishes; this is typical of his level of investment.

At Anthropic, we discovered AgI (Iodargyrite aka silver iodide) years ago and stuck it in this box

I think Clio is great

1) Scheming emerges if models "really care" about something 2) Claude 3 Opus really cares about not being harmful IMO it's mostly a paper about *scheming*, and "alignment" is a muddying frame here.

Nice paper. Also a good opportunity for me to explicitly admit that I was wrong about the distraction argument. (To be clear, I didn't change my mind yesterday because of this paper; I did so over a year ago and have said so on talks and podcasts since then.) There are two

Opus 3 is a very special model ✨. If you use Opus 3 on the API, you probably got a deprecation notice. To emphasize: 1) Claude Opus 3 will continue to be available on the Claude app. 2) Researchers can request ongoing access to Claude Opus 3 on the API: support.anthropic.com/en/articles/91…

IMHO this is an *ideal* way to use LLMs' own words directly in your own writing, which usually is bad! For functional sections that *aren't* an expression of your ideas and taste — such as a summary of someone else's work — let Claude do that part, and put it all in italics.