Grok Bot, Muse, and Dots Shipped the Same Agent
xAI, Meta, and OpenAI shipped always-on agents 49 days apart with the same architecture, but none of them measures whether the promised learning loop actually improves.
ARI Bench
Articles & analysis
Reporting and analysis on recursive self-improvement, frontier models, and AI research.
RSI Is Coming Faster Than Expected, and the Best Rogue-Model Plan Scores 3 Out of 5
What frontier labs disclose about containing rogue models, and how that compares with their timelines for recursive self-improvement.
The Insiders Disagree by Orders of Magnitude
Six insiders on recursive self-improvement: their claims, the numbers, and where they disagree.