TechFlow news, on January 22, according to Decrypt, Anthropic CEO Dario Amodei said at the World Economic Forum in Davos that artificial general intelligence (AGI) could surpass human performance in most tasks within one to five years, potentially achieving superhuman capabilities as early as 2026 or 2027. Amodei noted that AI has already begun automating the creation of itself.
Anthropic's internal engineers now primarily rely on models to write code. Google DeepMind CEO Demis Hassabis took a more cautious stance, estimating a 50% chance of achieving AGI before 2030. Both executives agreed that entry-level white-collar jobs will face early disruption, with up to half of professional entry-level positions potentially disappearing within the next five years. They also warned that governments may be underestimating the speed and scale of economic and geopolitical risks brought by AI advancement.
