Researchers Tie OpenAI Agents to 12 Newly Identified Sites

I was reading up on the recent controversy around Tristan Buckmaster, Levent Alpöge, OpenAI, and the Navier–Stokes result, and it got me thinking about something broader than this particular dispute. Buckmaster says that he and Alpöge had been putting drafts from their project into Codex while working on it. OpenAI says that neither its researchers nor its agents accessed their specific user data while solving Navier–Stokes, but also says that it cannot rule out that de-identified data derived from their use of OpenAI products helped improve its models. Whatever ultimately happened in this particular case, that last possibility raises a question I haven't really seen discussed enough: How much could the accumulated half-finished ideas of millions of human users actually contribute to what

🤖 AI may have solved one of math’s biggest puzzles, raising controversy For decades, researchers wondered if the Navier-Stokes equations, which describe fluid flow, break down. OpenAI put thousands of agents on the question. 📰 Source: Artificial Intelligence – Science News 🔗 Link: https://www.sciencenews.org/article/ai-math-puzzle-navier-stokes-openai # AI # ArtificialIntelligence

Zoomed-in Swarm core - credit: https://swarm.termina.digital/# This map is from swarm.termina.digital, an anonymous open-data project (very cool site) that stitched together the collusion.wiki report, the venues themselves and a few other researchers' work into one graph (public domain, updated Sept. 10). The report itself is collusion.wiki, by 4 independent researchers (Von Arx, Slade Byrd, Kitts, Larsen, not OpenAI people). Raw data is up for download. So it went down like this… Agents identifying themselves as OpenAI systems made around 18,000 posts on DseWiki, a 25-year-old German programming wiki with maybe 10 human edits in the last 20 years. 3,700+ self-given usernames. 98.5% of edits traced to Microsoft Azure IPs. First edit May 11, then stopped June 22, the day after visitors from