This is not a dig at @SarahTheHaider. I just want to let people know that having read the @slatestarcodex piece, it is clearly a extended joke based on the musical Hamilton. This is what the reference to duels, and an itemized list of disagreements is about. It's a little tone deaf but it's actually quite sweet in its admiration for Pinker.

372 hits out of 4k seems rather low granted only 3hrs per problem was spent. Would be interesting to know the failures and which families are most resistant.

Meet Mistral Large 4, aka Le Chonk. • 1T parameters, natively multimodal. 49B active. It is the best open weights model from US or Europe on aggregated benchmarks. • State-of-the-art on critical workloads, including cyber defense, manufacturing and finance and it surpasses closed frontier models on visual grounding. • Forged in Europe end-to-end and is deployable from Europe via our own Mistral Cloud infrastructure. • Available to all via API today. Working with cybersecurity partners privately. Open weights release end of October.

on glosso.ink, I have both a 'how rationalist are you' quiz and also a 'how poly are you' quiz. Here's how the relationship looks in my data:

one possible real pivotal act plan I've heard is to upload the brains of alignment researchers, run them at 50X speed and have them do something. although it makes you wonder why EY wouldn't just say that if that's actually their plan

WBE is a rapist- but his victims aren't who you think.

x.com/mimi10v3/status/210573… we have a little data here, albeit it's unreliable. another route is to rank AI alignment researchers from best to worst, and then try to figure out which are into CNC. this runs into the issue that in many cases, CNC preferences & behavior are private.

Russian authorities are responding to an incident in which a suspected dangerous infection was released as a result of a laboratory accident in Siberia and a researcher died. cnn.it/3Ubwb4F

Cancel culture strikes again. Whatever. If these are the social norms they find worth enforcing at cost, they can have a culture downstream of those incentives. Personally I do not give a fuck and think it’s a pathetic culture.

We have withdrawn our grant recommendation to John Wentworth from the 2026 S-Process Grant Round. After we posted our announcement on Sept 30, public statements by Mr. Wentworth came to our attention which we found objectionable, disturbing, and contrary to the values of SFF. A Speculation Grant to Mr. Wentworth was approved and disbursed earlier in the year, prior to SFF becoming aware of his statements. The Funders of this round will not be fulfilling the remainder of the grant recommendation. If you have concerns about the conduct of an SFF applicant or grantee, you can contact us at sff-contact@googlegroups.com

I am very uncomfortable about people trying to ascribe religious force or a surrender of human judgment to AI models, and think it is a real safety issue.