Follow the evidence.
Understand the change.

A current, evidence-linked intelligence platform showing what changed in AI, what is credible, what is disputed, and what people should actually use.

50+

Sources tracked

Daily

Briefing published

Public

Corrections ledger

What we track

Frontier, open-weight, and local models. Pricing, capabilities, hardware requirements, and independent benchmarks.

Methodology-aware comparison. Vendor claims vs independent results. Saturation, contamination, and reproducibility flags.

Papers, repositories, releases, and community discoveries with evidence grading.

Confirmed Disputed Unverified Vendor-reported Corrected Outdated

Every rating is decomposable into source class, evidence quality, corroboration, timeliness, and independence. Read our full methodology →

Five things that matter today. Model releases, pricing changes, important research, security issues.

Get the daily briefing by email or follow via RSS. Briefings published as downloadable Markdown.