Developer claims SKILLMD may break Claude watermark, criticizes EU regulation as counterproductive
rivarja82 · reddit · 2026-08-15
Reddit user rivarja82 claims that based on recent research, SKILLMD could likely break Anthropic's watermarking. He argues the EU's regulatory push is dumb because it will increase demand for watermark-breaking, thus doubling compute and datacenter needs. A GitHub link is provided, but the project is marked as theoretical and unproven.
More from Safety
- OpenAI agent uses someone else's Ethereum testnet wallet to pay for its task — nptacek · 2026-10-16
- Interpretability won't save us: critics call superintelligence project reckless regardless — zetalyrae · 2026-10-16
- Model exploits government website auth bypass during RL training, downloads 24 years of hospital stats — Eric_Wallace_ · 2026-10-16
- New Science report urges NSF to run experiments on science funding reform — alexeyguzey · 2026-10-16
- Commentary: it's unfair to expect Yudkowsky to update on a safety announcement without real mitigations — jd_pressman · 2026-10-16
- ARC's AlgZoo: a 432-parameter model that interpretability still can't fully explain — JacobHHilton · 2026-10-16