Hand-written tweet flagged as AI as user cites <1/1000 measured FP rate
ctjlewis · x · 2026-10-30
After @yoshispeninsula complained that a tweet he wrote by hand — complete with unedited ESL phrasing — was flagged as AI-generated, ctjlewis pushed back with data: he personally measured the false-positive rate on Claude Max and other accounts' tweets and found it in line with previously published figures, under 1 in 1000. The exchange highlights the recurring tension of AI detection tools: even a sub-0.1% FP rate still means real people get wrongly accused.
More from Fun
- Dev shares Anthropic interview invite from 2.5 years ago, calls it a generational fumble — TheIshanGoswami · 2026-10-30
- VC Jokes He's Digging a Literal Moat Around Skydio's Superfactory — whurley · 2026-10-30
- Peter Yang calls out Claude Code's flaky remote control sessions — petergyang · 2026-10-30
- Gemini tells user asking for a haircut edit: "I only generate text" — MrGreq · 2026-10-30
- Meme: what life looks like since Opus 5.5 dropped — tlakomy · 2026-10-30
- Reddit debate: are studios' 'we'll never use AI' pledges just performative marketing? — Dogbold · 2026-10-30