TLDRocket
Sign in

AI Evaluation

360 summarised stories about AI Evaluation, each linking back to the original source. Browse all topics →

+ Follow this topic

Sunday, 13 September 2026

Brand New AI Solves a Millennium Prize

Zvi (Don't Worry About the Vase) 1 month ago 16 ● 21 sources

OpenAI’s internally run AI agent efforts led to a reported solution of the Millennium Prize problem on Navier–Stokes and triggered disputes about what was learned, verified, and who should get credit. The effort solved Navier–Stokes in 88 hours, after an internal run that used about 130 billion output tokens for Lean formalization and verification taking an additional 17 hours for Astra. The aftermath shifts the focus to how AI-assisted mathematical research was produced and accounted for, including resource use and contested authorship and data-use claims, rather than to prize eligibility.

It passed CI. It passed your evals. The customer still got the wrong answer.

The New Stack 1 month ago 46

The article describes how AI agent deployments can pass CI and automated evals while still producing wrong answers for customers, leaving teams with an “observability gap” where failures can’t be alerted on as “wrong.” The Dynatrace 2026 State of SRE and Platform Engineering report, based on 919 enterprise leaders surveyed globally, found that 77% embed observability in some services but only 40% fully integrate it across all deployments. It argues that teams must add run-time evidence beyond traces—especially retrieval inputs/filters, tool-call/tool-result trajectories, and linked evaluation results—so missing retrieval preconditions and other causes can be reproduced and fixed with regression tests and version-validity checks.

The daily briefing

Every AI story that matters, in your inbox by 8am.

TLDRocket reads all relevant sources, removes duplicate coverage, and summarises the day in two minutes. Follow companies and topics for alerts, or get the briefing in Slack. Free, no spam, unsubscribe anytime.