Karpathy, Who Called Today's AI Agents 'Slop,' Joins Anthropic to Use Claude to Build the Next Claude

Andrej Karpathy, the Tesla AI veteran who recently dismissed the current wave of AI agents as "slop," has joined Anthropic in a move that suggests the field's most exacting minds believe Claude itself holds the key to building smarter AI systems. The hire signals that Anthropic is betting heavily on a recursive loop of machine self-improvement—using its existing model to architect and train its successors rather than relying solely on human guidance. For an industry struggling to move beyond incremental gains, Karpathy's arrival with such a provocative mandate raises a sharp question: is this the path to genuine AI progress, or a cautionary tale about systems optimizing themselves beyond human comprehension?

Tech Times·May 31