Yale economist Pascual Restrepo argues that AGI will skip automating most jobs because the economic incentive simply isn't there.

Restrepo's NBER paper challenges the assumption that AI will automate all work it's technically capable of doing

The key insight: AI adoption depends on cost-benefit economics, not just technological capability

Most human jobs lack sufficient economic value to justify the investment in automation infrastructure

The debate shifts from 'what can AI do?' to the more important question of 'what will AI companies actually bother automating?'