Turing Award winner Richard Sutton argues that conventional generative AI systems cannot perform genuine scientific discovery because they lack built-in evaluation mechanisms to assess their own results.
Sutton identifies a fundamental limitation in current generative AI: without the ability to evaluate outputs, these systems cannot sustain scientific breakthroughs. Novel discoveries emerge only momentarily before disappearing, he explains.
According to Sutton, AI systems like AlphaGo and AlphaProof demonstrate what's required for genuine creativity in science—integrated feedback loops that allow systems to test and validate their own findings. This evaluation capacity separates truly creative AI from systems that merely generate plausible-sounding but unverified outputs.
The distinction matters for fields where verification and reproducibility are essential. Generative models excel at pattern recognition and synthesis but lack the iterative testing mechanisms necessary for scientific validation. Sutton's assessment suggests that achieving AI-driven scientific discovery will require architectural changes beyond current generative approaches, incorporating built-in evaluation and refinement capabilities.
Two former Department of Government Efficiency staffers have launched a new company to acquire businesses and slash operational waste using artificial intelligence, extending DOGE's cost-cutting playbook to the private sector.
Gary Marcus, NYU emeritus professor and AI entrepreneur, described a recent executive order on artificial intelligence regulation as a watershed moment for U.S. policy, marking a sharp departure from the previous administration's deregulatory stance.
Microsoft released MAI-Thinking-1, a new AI model family designed to improve reasoning capabilities. The launch includes seven distinct model variants targeting different use cases.