Lately, I've been tinkering on an AI Doompop annotations site (here).Last night, Claude notified me of Hello World, AI Doompop , which inspired me to share my project on LessWrong.Features include:View annotations as a music video plays:View annotations alongside a selection of AI music videosExplore a lineage of individual songs:The lineage tree around "Claude-Pop -…

Epistemic status: Mix of exploratory and my best guess. I do try to substantiate my claims with relevant research where possible. TLDR/Tweet summary á la Emergent Ventures: Crying maths professor = AI meaning crisis is here. Solution: duolingo for classical keyboard improvisation (Bach-Mozart) but it actually works and 2xs learning speed by combining passive haptic feedback learning…

A three-year AI pause could buy 1000 years' worth of AI safety work. See: https://1000yearsofaisafety.comHow much AI safety work a pause produces, in years of the whole 2025 AI safety field's work. The government bans new large training runs and gets the top three US labs to redirect their GPUs and researchers to safety research.We…

Graham Greene described G. H. Hardy’s 1940 book, A Mathematician’s Apology, as the best account of what it is like to be a creative artist. Hardy had many friends, wide-ranging interests that included the game of cricket, and empathy for the less well-off. But it was his creative work in pure mathematics that gave Hardy…

We've written a paper on optimized multi-agent attacks, for which we present a brief writeup in this post. See the paper for more detail. This work was also presented as a spotlight research talk at the MATS Summer 2026 Symposium here. Tl;dr Recent AI cyber-related incidents exposed vulnerable monitoring surfaces to detect attacks stemmed from…

In July 2026, Korean Go player Shin Jin-seo defeated KataGo 2-1, with only[1] a 2-stone handicap (favoring Shin). I thought this was very surprising, and posted on LessWrong and Twitter asking how to interpret this. I didn’t get any entirely satisfactory answer, so I decided to spend ~one day looking into it and writing up…

I believe the flow of interesting things to people who would be interested in them is not currently as effective as it could be. I also believe (a) that such a system could be built and (b) it would have important implications for society.I will now attempt to describe what such a system could look…

I am Austrian and wrote this in German. The English is a literal machine translation, not rewritten.Everywhere now, catastrophic harm to humanity is being discussed, and various scenarios. But the other percentages are not mapped in detail by outcome, the way it is done with nuclear scenarios, where people look at this more closely. That…

IntroductionI recently watched the video "'AI Existential Risk' Was Invented to Distract Us" by Adam Conover of Adam Ruins Everything where he expressed his views on why he believes the claim of the extinction threat posed by AI is being used as a diversion from other harms caused by AI. I’ve seen this narrative being…

TL;DR. We run the CoT-Control evaluation suite on GLM, Kimi and DeepSeek models and find that while Kimi and DeepSeek models have poor CoT control, GLM models are outliers in their ability to control CoT. We extensively investigate this behavior in GLM-5.1 and show that previous generation GLM models (GLM-4.5 and GLM-4.7) also have similarly…