Google DeepMind CEO Demis Hassabis has predicted that artificial general intelligence (AGI), defined as a system capable of exhibiting all human cognitive capabilities, could be achieved within five to ten years. Speaking at a London event, Hassabis emphasized the challenges in developing AGI, including understanding real-world contexts and overcoming technical limitations. The announcement coincided with Google DeepMind's expansion in the UK, including the introduction of Agentspace data residency and the addition of Chirp 3 to Vertex AI. These efforts aim to bolster AI adoption among UK enterprises, supported by £280,000 in cloud credits for startups and expanded AI training programs. The UK government is also promoting AI development through regulatory changes and investment in regional AI zones. Google DeepMind has also seen significant organizational growth, with its workforce more than doubling to 5,600 employees in two years, including a team of 1,800 working on the Gemini chatbot. This expansion positions it to compete with rivals like OpenAI and Anthropic in advancing AI capabilities.
While I appreciate this study, I'm also a bit worried its headline result is misleading—it only measures performance on a narrow set of software tasks. As of March 2025, AIs still can't handle 15-minute robotics or computer-use tasks, despite what the headline plot might suggest. https://t.co/idkZNJpL21
BRILLIANT Paper. 💡 Practical Meaning If this pace holds, AI might soon handle larger coding or research tasks without constant oversight. 📈 Observed Trend Researchers compare model success rates on tasks against the hours expert humans need. The longest tasks current models https://t.co/U8mKNRyv6W
Measuring AI Ability to Complete Long Tasks https://t.co/jqwleTZpvN https://t.co/OadL5Stxvt