Anthropic researcher expects AI to match top experts by 2027
Tao Lin, a technical staff member at Anthropic, predicts that by the end of 2027, AI could match or beat top human experts in coding, filmmaking and other fields after about 1.3 more years of development. He works on Claude model safety, including automated alignment, AI control and safeguards.
Lin warns that AI's impact on jobs may unfold step by step. AI takes over current work first, workers move to new roles, and those roles may later be automated too. In his scenario, almost everyone could lose jobs long term and rely on government support.
Anthropic research shows 68% of Claude usage involves tasks AI could theoretically do alone, but that does not mean jobs will be replaced. Limited robot adoption makes mass unemployment by 2027 very unlikely. Replacing physical work still needs advanced robots, automated factories and supporting systems.
Comments
Loading comments…