@METR_Evals Anonymous feedback: https://xitter.cf/t.co/0k6Duym4fI
Berkeley, CA
Joined August 2022
- Tweets6.8K
- Following4.2K
- Followers2.1K
- Likes51.1K
Pinned Tweet
Hi all, I'm interested in feedback!
You can leave anonymous comments here: admonymous.co/jsd
Personal news: after two years at @EpochAIResearch, I’ve joined @METR_Evals to work on embedded assessments of AI R&D acceleration: going inside AI developers to measure how much AI is speeding up AI development, and publishing our findings.
I’m very grateful to everyone I worked with at Epoch. As I leave, I especially want to thank @merilalama, Epoch’s COO. Operations work too often goes unsung, but her judgement and hard work are a huge part of why Epoch has accomplished so much.
Both orgs are hiring. If you want to help the world understand where AI is headed, consider applying to Epoch epoch.ai/about/careers or METR metr.org/careers
Ok, Berkeley Bowl is great.
Also, over the course of a week, 3 people I know independently compared visiting Berkeley Bowl to Russians crossing the iron curtain and discovering the abundance of Western supermarkets.
2 of them mentioned this photo of Yeltsin:
I know very little about macroeconomics, have been trying to understand a bit more, and a tentative takeaway is that a more than I expected is justified by frictions and sticky prices.
jsd retweeted
Language models are getting much, much harder to fully understand with scale, not easier! (Thread)
Hey @nickcammarata, have you interpreted my 432-parameter model yet? Do you think you'll manage it by year end? That's 4 hours per parameter
