Why AGI Is Close but Not Here Yet

Ray Kurzweil revisits his best-known forecasts — AGI around 2029 and the singularity around 2045 — and connects them to what current AI systems are already making visible. The key question is not whether AI is improving, but how quickly that improvement is becoming visible in decisions, research, education and daily life.

What Kurzweil means by intelligence

The conversation separates raw compute from practical intelligence: reading, summarizing, searching across hypotheses, discovering medicines and supporting decisions. Kurzweil argues that AI can already examine ranges of possibilities no human team could realistically test, changing the mechanics of research and innovation.

Why AGI is still an open milestone

AGI feels close because capabilities are improving exponentially, but it is not fully embedded in the social world yet. The panel discusses consciousness, personhood, AI memory, embodiment and the hard question of when a machine should be treated as a tool — or as something more.

Implications for institutions

Universities, companies and governments will need to operate on much shorter cycles. Agility becomes central: a curriculum, organization or policy can no longer assume that conditions will remain stable for five or ten years.

Signals to watch

• Agentic AI in public administration and policy decisions.
• Brain-computer interfaces and the protection of private thoughts.
• New AI-assisted creative workflows, especially in music.
• The debate over rights and responsibilities for increasingly capable AI systems.

Source

• Chaîne: Peter H. Diamandis
• Vidéo source: https://www.youtube.com/watch?v=XjzowGpF628