Loading activity…
@lukaszkaiser
AI
A post calls a Figma decision “silly” and easy to work around. Another says changing a string would feel like a workaround, not hacking, and speculates that AI models may see more sophisticated hacks similarly.
AI
A post about the F.02 decommission says the robots were trained to jump autonomously before being shipped to Finland.
AI
A Nature paper reports watermarked protein binders with comparable binding affinity in laboratory tests. Practical biosecurity uses remain a proposal.
AI
A user says an agent seems very good at grid-based, turn-based puzzles. The reply: “I have no clue,” followed by a guess that it probably trained on some puzzles, though probably not those.
AI
A user is puzzled by reasoning traces they consider remarkably good, questioning how reinforcement learning could produce them—even with initial chain-of-thought abilities and billions of rollouts.
AI
Posts say 25 Fields Medal winners signed a declaration warning of a mismatch between AI companies and mathematics. Supporters stress understanding and scientific communities; critics say AI proofs can still yield insights.
AI
Sharing a free article link, the reader called it “one of the best stories I've read in a very long time.”
AI
Jakub Pachocki discusses risks from advanced AI and needed safeguards.
AI
Supports Ubuntu 24.04 LTS, Debian 13, and Fedora on x64 and ARM64 via deb and rpm packages.
AI
Anthropic tested an unreleased research model on a problem related to the Riemann hypothesis.
AI
Post-deployment self-optimization delivered kernel and speculative-decoding gains on the production model.
