This is the weekly visible open thread. Post about anything you want, ask random questions, whatever. ACX has an unofficial subreddit, Discord, and bulletin board, and in-person meetups around the world. 95% of content is free, but for the remaining 5% you can subscribe here. Also:
1: You might remember TracingWoodgrains winning the 2019 SSC adversarial collaboration contest with his piece on whether schools adequately served advanced students. Six years later, Trace (real name Jack Despain Zhou) and Lillian Tara are starting the Center For Educational Progress, a think tank to promote their agenda (mostly ability tracking). Read the manifesto here. If you’re interested in volunteering, following along, or helping with funding, check out their Discord server.
2: Google DeepMind’s alignment team has new job postings, for research scientist and research engineer, available “in London, Zurich, New York, Mountain View or San Francisco”. Applications open until 2/28. They’ve also released their own free short video course on AI safety.
3: Correction to the Cruz woke science article: I conjectured that unrelated science grants contained a sentence about women and minorities to please the Biden administration, but even that was granting the Trump narrative too much. Commenters pointed out that grants being judged on the “broader impact criteria” - a seven pronged list including outreach and benefit to women/minorities - actually dates back to 1980 and neither Biden nor the current round of wokeness was involved. And not a correction, but a clarification - several people suggested that even 40% of grants being “woke” was bad. The article didn’t intend to claim that 40% of NSF grants were woke - only 40% of the NSF grants that Ted Cruz and the Commerce Department had previously identified as woke. Those in turn are about 5% of all grants, so (assuming Cruz didn’t have false negatives) only about 2% of total grants were woke.
4: Sorry, I had to briefly delay announcing the winners of last year’s Metaculus/ACX Forecasting Contest while some scores were recalculated. Here are the ones who responded to my email (if you didn’t get it but should have, email me at scott@slatestarcodex.com):
ErrorMargin is a quant trader. He says he has "only dabbled in predictions before, but did this round with a friend and put way more effort into explicit modeling". He has a new blog (with one post) at errormargin.com, and says "I'd love to receive any emails from ACX readers to me@errormargin.com. I'm always happy to hang out with fellow forecasting nerds or rationalists in and around London."
Lasse Schettlinger is from Germany and is working on a masters in Quantitative Economics. He currently finances his studies by working as a croupier.
J is an equity options quant in London. He has a retrospective on the 2024 contest, along with his models/predictions for 2025, on his blog. He say he is "unknown enough that I actually enjoy getting emails from people at j@thedissonance.com", and is open to meeting new people in London or hearing about job opportunities in quant dev / research or software.
