Do VLMs Plagiarize? UC Berkeley's Jitendra Malik Weighs In on Credit Assignment in the Astra Demo Debate
JitendraMalikCV · x · 2026-09-19
- Junior researcher Toru argues that Astra's impressive dexterous manipulation demos expose two unanswered academic questions: how to assign credit when an agent synthesizes many prior works into a new result, and how academia can retain talent when incentives hinge on credit.
- UC Berkeley professor Jitendra Malik agrees: while VLMs like Astra make prior research far more accessible—like encyclopedias and Google Search before—society historically relied on patent disclosures and citations for credit assignment. Should models that don't cite sources be accused of plagiarism? He notes his own co-authored in-hand rotation papers may have been used in Astra's pen-spinning demo without attribution.
- He also references an open letter from 25 Fields Medalists responding to the Navier-Stokes result, complaining about disruption to mathematics research.
More from AGI Musings
- "Understanding" is a weak descriptor: Lean proofs are precise understanding, and machines build on them — sytelus · 2026-09-19
- Your agent becomes the storefront: shopping without opening a single tab — armand_ruiz · 2026-09-19
- 'Post economic' becomes SF dating slang, a sign of AGI-era wealth anxiety — signulll · 2026-09-19
- Alignment researcher criticizes field's fixation on near-term ASI 'foom' — tszzl · 2026-09-19
- 500 LLM agents ran a full propaganda campaign on a simulated X with zero human input — ziv_ravid · 2026-09-19
- David Chalmers and Robert Long spotted together at ConCon conference — RosieCampbell · 2026-09-19