Prediction◆Article
Intology's Locus harness under-elicits today's AI systems' ability to automate AI R&D, and the human baseline on PostTrainBench v1.1 (51.1%) will likely be exceeded before the end of 2026.
Locus's performance jump suggests we are under-eliciting AI systems for automating AI R&D; the author predicts the human baseline on PostTrainBench will be exceeded before end of 2026. ✦ AI generated
Jack Clark · Import AI · 2026-08-10 · original ↗
Posts like this highlight how we are under-eliciting today's AI systems for their ability to automate AI R&D - especially striking is how the company can jump the performance of Opus 5 by 10 absolute percentage points with a better harness. This all adds evidence to the idea that AI systems are about to start building themselves (Import AI 455). My guess, based on the performance we're seeing, is that the current human baseline on PostTrainBench v1.1 (51.1%) will be exceeded before the end of 2026.
Read full article ↗excerpt · fair-use quotation
- ·Under-eliciting AI systems for automating AI R&D
- ·Simple better harness jumps Opus 5 by 10 absolute points
- ·AI systems about to start building themselves (Import AI 455)
- ·Human baseline 51.1% likely exceeded before end of 2026
Around this claim