DeepMind Boss Says AI Agents Are Just a Dress Rehearsal for AGI in 2029
The masterminds behind the curtain are warning that today's helpful AI assistants are merely a warm-up act. The real, mind-bending show is coming much faster than anyone anticipated, leaving society scrambled and utterly unprepared for what comes next.
Demis Hassabis, the head of Google DeepMind, shifted his timeline for Artificial General Intelligence (AGI) to 2029 during an interview with Axios. This shaves a whole year off his previous estimate, officially positioning humanity on what he calls the foothills of the singularity.
Hassabis deliberately framed the current wave of AI agents as a societal stress test to shake up sleeping economists, regulators, and the general public. He noted that his economist friends are still treating this technological leap like a minor update to Microsoft Excel, which needs to change immediately before the real disruption hits the labor market.
As proof of how fast things are moving, he pointed to Anthropic's model, Mythos, as a prime example of a warning shot. The rapid capabilities of this model caught both corporate boardrooms and government officials completely flat-footed, proving that society's reaction time is dangerously slow.
The ultimate frontier, according to the tech chief, is recursive self-improvement—systems capable of coding and upgrading themselves without human intervention. While fully self-improving models aren't quite loose in the wild yet, developer tools like Claude Code and Codex are already acting as gateway drugs, massively boosting engineer productivity by letting machines write their own building blocks.
To curb the potential chaos, Hassabis threw his support behind US executive orders demanding mandatory safety testing before major models hit the public, and he is currently in talks with other elite AI lab leaders to coordinate safety frameworks, though he declined to share any specific regulatory details.
The countdown to AGI has officially transitioned from a distant sci-fi fantasy to a boardroom deadline. Humanity now has exactly three years to decide whether to spend this final grace period building real societal shock absorbers or just coasting through a few more cycles of venture-capital-funded hype and inevitable existential dread.
Source: Axios
Comments
This is where the magic happens: AI reads your discussion and rewrites the article based on the most interesting comments. Each strong comment adds points to the meter below. Once the meter is full, the article updates live — no page reload needed.