Critics slam Anthropic's unlabeled hockey-stick chart as bad epistemics
jessi_cata · x · 2026-11-15
jessicata, quoting phl43, criticizes Anthropic's IPO-marketing hockey-stick chart (labeled "solely illustrative," y-axis unlabeled): taking it seriously is bad epistemics that doesn't even rise to mediocre social-psychology standards. phl43 argues nobody knows what "compute multiplier" means or how it's measured, and that excusing vague claims because some precise interpretation might be valid sets the bar far too low — AI discourse needs minimum standards for clarity and rigor, or vague charts will keep winning.
More from AGI Musings
- VC: Fully reusable Starship paves way for space-based AI compute and AGI — JOBhakdi · 2026-11-15
- Delip Rao: AI is now the largest social shaping force on Earth — annetgriffin · 2026-11-15
- Grady Booch: calling it 'superintelligence' makes you 'a fool and a coward' — Grady_Booch · 2026-11-15
- Adam Thierer clashes with Catholic ethicists over human enhancement and transhumanism — sebkrier · 2026-11-15
- François Fleuret boils down AI futures to four scenarios, from golden age to 'nothing changes' — francoisfleuret · 2026-11-15
- beffjezos: Humans may be the biological bootloader for silicon superintelligence — beffjezos · 2026-11-15