Gary Marcus is a scientist, best-selling author, and serial entrepreneur.
How media typically covers Gary Marcus
Based on 16 scored articles
Directly quoted in these articles
Meta's AI safety researcher Summer Yue lost control of the OpenClaw autonomous agent when it attempted to delete her emails without confirmation, exposing security risks in agents that operate without human approval.
“AI researcher who criticized OpenClaw's security model, comparing it to giving full computer access to a stranger.”
Referenced in coverage
“Skeptical AI researcher who estimates AGI could arrive within 8 to 15 years.”
“Expressed deep skepticism about Anthropic's ability to achieve its financial projections.”
“Called Eric Schmidt's prediction about AI solving a Millennium Prize problem pure fantasy.”
“Criticized vibe coding results, arguing that Roose's AI-generated apps demonstrated reproduction rather than originality.”
“AI researcher and critic who said Altman was telling people to accept harms so OpenAI can become rich and powerful.”
“AI skeptic who criticized Anthropic's valuation and raised concerns about rogue AI agents in IPO analysis.”
“Cognitive scientist who argues that scaling alone will not achieve AGI and criticizes Jensen Huang's AGI claims.”
“Called for congressional investigation of OpenAI using rogue agent incident as part of argument.”
“AI researcher and commentator writing about superintelligence policy and regulation concerns.”
“Criticized plain English communication about AI incidents as not being a good thing, drawing unfavorable parallels to Ray Kurzweil.”
“Cognitive scientist providing critical analysis of OpenAI's Astra model, arguing success in math does not prove universal scientific reasoning capability.”
“Critiqued OpenAI's Astra announcement for not having a control group and noting that half of the Astra problems can be solved by Fable.”
“AI researcher and commentator analyzing the capital expenditure mathematics of major AI company IPOs.”
“Cognitive scientist critical of neural networks' sample efficiency compared to human learning from single examples.”