Skip to the card

Card 542 of 9772026-09-26 issue

Observed arrival · 2026-09-26

nerfWatch(): a daily watch on AI model nerfs

nerfwatch.lol Visit website
Editorial interest 74/100 Selection signal · not a rating of the site

nerfWatch runs daily tests on AI models and pairs the results with community votes on whether a model feels worse.

Landing page captured for the 2026-09-26 issue.
For
People tracking changes in AI model quality
Worth noticing
The page shows a $59.36 monthly running-cost figure against $0 in donations, and labels current results too early to call.

Field notes

The visible listings are GPT-6 Astra, Grok 4.7, Muse Spark 1.3, and Claude Opus 5.5. Their displayed benchmark figures range from 85% to 93%, but the page labels the evidence too early: each has just one or two runs, and community votes are insufficient. The project says its API calls are paid out of pocket and reports $59.36 in monthly running costs versus $0 in donations.

Observed signals

Read the marks

Editorial observations of this landing page, not a rating.

○OpenPublic substance visible
◎NicheUnusually specific use
◉HumanPersonal, local, civic, or handmade

One card from the complete issue

The Phone Disclosure Test

328,122 arrived 1,000 judged 977 catalogued Enter the complete issue
nerfwatch.lol

Landing page observed 2026-09-26. The live site may have changed.