AIAI Snake Oilabout 2 hours ago
AI agents can't yet do open-ended AI research
7 min read
Frontier AI agents failed open-ended AI research in a new shadow evaluation by Princeton researchers. Expert authors rejected both agent papers after six days and thousands in API credits. Agents lacked judgment, ignored instructions, and failed to backtrack or spend budgets.
Level
Hype check
Tap to vote and see what everyone thinks.
Summary by ByteBrief
