Dev Calls Default AI Text Watermarks Unfair, Ships Simple Unmark
haltakov · x · 2026-11-20
Developer @haltakov argues that watermarking AI-assisted text by default is unfair: something you wrote yourself then polished with AI gets the same mark as slop generated with barely any human input—the watermark doesn't reflect how much thought went in, and that choice should be yours.
He built Simple Unmark accordingly: paste text and it removes hidden watermark characters, then rephrases it with a different model to defeat detection. It's built on @guillaumemeyer's open-source watermarks-remover library—another move in the escalating game between AI provenance watermarking and counter-tools.
Related event: Simple Unmark Launches to Strip AI Text Watermarks(4 posts)→
More from Safety
- 2M canary tokens reveal 3 breached AI providers and resellers stealing prompts — dyn___ · 2026-11-20
- Perplexity adds Decider to Context API to score pages and catch prompt injection — denisyarats · 2026-11-20
- California panel report targets RSI risks: compute disclosure, SL3 security, outcome standards — gleech · 2026-11-20
- Hacker Swapped API's DeepSeek for Abliterated Version, Undetected for Two Days — moyix · 2026-11-20
- Claim: METR is an Anthropic-linked EA arm, not an independent evaluator — gerardsans · 2026-11-20
- Fuck Your AI Safeguards: a hands-on look at bypassing AI guardrails — speckx · 2026-11-20