Researchers as 'meat relays' for Claude can't answer questions without it
ChongZzZhang · x · 2026-11-21
ChongZzZhang criticizes a new academic phenomenon: some researchers act as mere 'meat relays' for agents — supervisors' questions become prompts to Claude, Claude's outputs become their progress, and they can't defend technical nuances when offline. He contrasts with those who work hard but skip publicity, letting papers sink among 9000 submissions in slow peer review.
More from Companies & People
- tenobrus joins METR, debuts major eval report on model capability gaps — morqon · 2026-11-21
- Ex-Datadog employee: we used to send donuts to teams having outages — that would bankrupt us now — Technovangelist · 2026-11-21
- Gary Marcus mocks Anthropic ad claiming Claude discovered gel electrophoresis — GaryMarcus · 2026-11-21
- Arcfront expands hiring, turns real hospital and factory work records into training data for frontier models — testingcatalog · 2026-11-21
- Alex Hauptmann, CMU LTI research professor for 42 years, dies at 67 — windx0303 · 2026-11-21
- Gary Marcus maps the AI bubble's 'Five Stages of Generative AI Grief', from denial to quantum — GaryMarcus · 2026-11-21