Developments elsewhere that bear on this problem: research, results, incidents and policy, each with its source. The UNSOLVED editors choose them every week, with AI research help. They are not verdicts: the review board rules only on contributions.
Explaining why it replaced Git with its own source-control system, Anthropic gives the rise in daily merges and attributes it to Claude's ability to write code. It counts output, not research speed, so it is not an acceleration measure, but it is a rare published figure on how much AI now works inside a lab.
Anthropic · Redesigning source control at Anthropic
Fitting an internal capability index across 123 models and 92 benchmarks, Google DeepMind finds the slope after 1 April 2026 (0.118 a month) not significantly above the slope before (0.091). It reports that most production code is now more than 75% AI-written, but 88% of pull requests still need human steering. A rare lab estimate of end-to-end acceleration.
Google DeepMind · Gemini 4 Argon Safety & Security Report
Epoch gave six models 11 real tasks from its own work, graded by hand. The two leaders, Claude Fable 5.1 and GPT-6 Astra, were reliable on well-defined coding and analysis; open-weight models were not. All fell short on implicit standards, experiment design and reading flawed results. Epoch concludes AI cannot yet replace its staff end to end. Small sample, one run per task.
Epoch AI · Can AI automate Epoch?
Epoch AI gave agents a Qwen3-8B setup with a recent self-distillation method (SDPO) removed and about 3,000 GPU-hours. Of the uncontaminated models, only GPT-5.6 Sol made an in-scope gain, about 15% of SDPO's (35% judged generously), close to earlier work; other scores came from out-of-scope selection, tuning or memorisation. One task, one run per model.
Epoch AI · Can AI automate AI R&D yet?