1.
2.
Arena, the AI leaderboard everyone uses, is now a $100M business
(techcrunch.com)
3.
Monitoring LLM behavior: Drift, retries, and refusal patterns
(venturebeat.com)
4.
General scales unlock AI evaluation with explanatory and predictive power
(feeds.nature.com)
5.
Ask HN: How are people doing AI evals these days?
(news.ycombinator.com)