Steering AI beyond automation
AI Snake Oil presents a research agenda and an action plan aimed at steering AI toward uses other than straightforward automation.
Related stories
Google says self-improving AI isn't anywhere close
Google's Haydn Belfield says DeepMind sees no measurable speedup in AI research driven by AI itself, with its internal General Capability Index showing steady rather than sudden growth. AI now writes more than 75 percent of production code, yet 88 percent of those changes still require human oversight.
Lab-in-the-loop at compute speed: Ian Quigley of Leash Bio
In a Latent Space interview, Leash Bio's Ian Quigley explains how the company mastered DNA-encoded library assays, a notoriously difficult technique, and used the resulting data to detect models that cheat. He also argues the same data is well suited to training AI agents to reason like chemists.
AI came for math. Now it's chasing the rest of science
MIT researchers tell Wired that AI could soon become the standard way to tackle theoretical problems in fields such as economics and high-energy physics. The piece frames this as the next step after AI's progress in mathematics.
Researchers use AI to design proteins that cut up other proteins
Scientists have used AI design tools to create proteins that cleave other proteins. The designed molecules can break chemical bonds that would otherwise remain stable for hundreds of years.
Datalab releases OmniParseBench, an open OCR benchmark with 16,288 pass/fail tests to audit OCR parsers
Datalab has released OmniParseBench, an open benchmark of 16,288 yes-or-no tests spanning 2,937 pages in 92 languages. Each score is a pass rate that can be broken down by tag, script or language, making it easier to see where OCR parsers fail.