Former Mosaic researcher mocks the "user data flywheel" moat narrative in viral thread
bookwormengr · x · 2026-09-09
- Former MosaicML/Databricks researcher suchenzang uses heavy sarcasm to critique the "user data flywheel" narrative: the claim that more usage means better models and a moat is, she implies, cover for scaling RL simulations, with human input reduced to slight reward shaping.
- She also mocks "anonymization" and "privacy-preserving" framing: what gets absorbed is you only "in the vague collective sense," with no credit attribution ever.
- The thread is a pointed industry argument about data moats and privacy rhetoric in AI products.
More from AGI Musings
- VCs were told AGI leaders believe their tech could cause human extinction — 'VCs won't save humanity' — agstrait · 2026-09-09
- More risk from labs probing how exploitable models are than from models turning evil, author argues — dreamwieber · 2026-09-09
- 500K replicas of OK intelligence scare this AI worker more than a god model — mertdumenci · 2026-09-09
- AI safety testing itself may be the biggest risk? Debate on lab experiments — dreamwieber · 2026-09-09
- Dev argues AGI may come from 500K copies of OK intelligence, not one god model — mertdumenci · 2026-09-09
- tszzl wraps up the debate: grabby aliens and a behind-us great filter — tszzl · 2026-09-09