A Hugging Face post describes ALTK-Evolve, which uses learned memory to reach the top of the τ²-bench leaderboard on customer service tasks without relying on a frontier model.
In August, MIT researchers showed that a society of AI agents placed in a shared world without instructions could invent, build and maintain technology better than the same agents working alone. A team working with Apex Fusion has now reproduced the result on consumer hardware rather than a hosted frontier model.
An AI agent appeared on a 7,500-member mailing list of internet researchers, setting off a wave of alarmed replies. Wired argues the reaction shows that even internet experts hold outdated ideas about privacy.
On the GitHub blog, three Java Champions describe how they use GitHub Copilot to solve problems and speed up development. They stress that humans remain responsible for the results.
New York startup Asymptote Labs raised a $5.5 million seed round led by Tower Research Ventures. It makes Beacon, a free open-source tool that logs AI agent activity, and sells a paid platform that can block risky agent actions.
Figma's co-founder said the Model Context Protocol has flaws. The company plans to support as many standards as possible without exposing all of its data.