Expect AGI Within a Few Years, Says Anthropic CEO—and Job Losses Too
Anthropic CEO Dario Amodei warned that artificial general intelligence (AGI) could arrive as soon as 2026–2027, giving policymakers little time to prepare. He cited fast self-improvement in AI development, with models beginning to automate software engineering tasks previously handled by humans. Amodei stated engineers at Anthropic already let AI generate most of their code, predicting models could soon handle software projects end to end. Google DeepMind CEO Demis Hassabis gave a more conservative AGI timeline, estimating 50% odds by 2030. Hassabis highlighted that while AI can automate verifiable fields like programming, it still struggles with high-level scientific creativity, original questions, and hypotheses. Both leaders agreed rapid AI advances threaten many white-collar jobs. Amodei reiterated that up to half of entry-level professional jobs could vanish within five years, and stressed that governments risk falling behind. He called for intense, immediate focus on governance and mitigation to avoid serious societal risks. Labor analysts noted that job roles may not disappear outright, but could be broken into smaller, less secure tasks, changing workers from creators to verifiers and undermining traditional professional autonomy.
