🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

The Exponentiating Mathematics (expMath) program BAA has dropped! sam.gov/opp/869c8d7351… Abstracts due May 15. Be sure to register on the submission website well in advance! (A few days ahead!)

There is no longer any not-super-biased sense in which any human is "clearly overall smarter" than o3, Claude 3.7 Sonnet, Gemini 2.5 Pro, etc. Nor are the AIs clearly overall smarter -- yet -- but notably, 2 years ago, there was no contest. Seems maybe notable or something???

You can say that again!

we’ve come so far since xkcd.com/251

Safeguarded AI’s proof checker will treat proof scripts as untrusted bytestreams. If “sorrys” exist at all, I think we should have a graded modal type system so that a term with sorrys only typechecks as a type that depends on a list of the UUIDs of those exact sorrys

AI researchers & Lean 4 pros: Want the details on DeepSeek Prover V2's recursive subgoal pipeline & RL? The full paper is freely available now. Go learn.

protocol, n.: a specified property of trajectories of an interface trajectory, n.: a mapping from a segment of time to the state of a system interface, n.: a subset of a system, which is simultaneously contained within the boundary layers of more than one of its subsystems

One of the simplest and most obviously worthwhile safety measures for powerful AI is a remote off-switch, and powerful robotics is no exception.