Anthropic's 335-page IPO filing flagged as AI-written by detector Pangram
andrewjdavison · reddit · 2026-10-27
Someone ran Anthropic's 335-page IPO filing through the AI-writing detector Pangram. A screenshot shows long passages on mechanistic interpretability flagged as AI-generated, while a paragraph defining "model weights" was marked human-written.
The irony writes itself — the AI-safety company's prospectus appears largely AI-drafted. Caveat: AI detectors are notoriously unreliable on formulaic legal text, so treat it as a fun data point, not proof.
Related event: AI Detector Flags Anthropic's S-1 Filing as AI-Written(2 posts)→
More from Companies & People
- Ashoka University launches one-year Master's in Applied Digital Health and Responsible AI — DrDatta_AIIMS · 2026-10-28
- Meta: 30% of PMs, designers and researchers now ship code, up from 5% a year ago — omooretweets · 2026-10-28
- Next Gen Maker Con, uniting Microsoft maker communities, opens registration — donasarkar · 2026-10-28
- METR researcher joining Anthropic sparks concerns over eval lab independence — ShakeelHashim · 2026-10-28
- Microsoft outlines 4 goals for Power Platform and Copilot Studio at PPCC26 — donasarkar · 2026-10-28
- VC observer: Anthropic is a 'pure expression' of venture capital practice — matt_slotnick · 2026-10-28