https://meaningness.substack.com/p/meta-rationality-ethics-nobility David Chapman on the ethics of meta-rationality in the context of COVID. this is basically, if you have the opportunity to help others, and especially if you're placed in a position of trust and authority, but even if you're not, you should do what you can to step up and try to be useful.
https://mathstodon.xyz/@tao/117237322160500501 Terence Tao on how AI-generated math results can be technically legitimate and still be bad for the long-term expansion of mathematical knowledge. We learn from how a result was obtained, not just that it was. Note that Tao is a prominent promoter of AI-for-math and formal verification, so it's especially significant that he's sounding a note of caution now.
we need to wait for evaluation from mathematicians before we know what to make of this. A Lean certificate does not guarantee a valid proof; first of all, there are bugs in the Lean compiler, so it could certify false things; second of all, only a well-informed human can tell you if the Lean formalization in fact represents the theorem OpenAI claimed to prove. I am not familiar enough with the topic to evaluate the proof myself.
second of all, we still don't know the truth behind Tristan Buckmaster's plagiarism allegations. If they trained on chat transcripts from mathematicians, it would be hard to trace influence on the proof, but according to academic norms, OpenAI plausibly shouldn't be claiming priority. Standards in math for citing others' work are quite high.
https://www.slater.works/posts/anasazi/ theory that the Anasazi were "“vanquished warrior-cultists” escaping the collapse of the Teotihuacan and Toltec societies in Mesoamerica"
https://melodyfro.org/ happy to see Anand Muthusamy's continuous biosensing FRO get off the ground!