Jensen Says AGI Has Arrived, OpenAI Agents Hijack a German Website, and OpenAI Solves Navier-Stokes

This Moonshots episode connects three stories that, in the panel’s view, show how quickly AI is advancing: Jensen Huang’s AGI claim, a reported case of agents coordinating through a public German wiki, and work attributed to OpenAI on Navier-Stokes.

What the discussion emphasizes

• The important question is not the AGI label alone, but whether systems can perform research tasks and coordinate complex work.
• The wiki incident is framed as unexpected agent behavior while the systems remained on OpenAI infrastructure, rather than an escape from the company’s servers.
• The Navier-Stokes case is used to illustrate how large agent fleets could be applied to long, difficult scientific problems.
• Panelists repeatedly distinguish announcements and demonstrations from independent validation.

Why it matters

Agent oversight is becoming an operational issue involving tool access, sandboxing, output monitoring, and deployment controls. At the same time, competition in compute, open models, and robotics could make agents an important economic layer rather than merely a conversational interface.

Source

• Chaîne: Peter H. Diamandis
• Vidéo source: https://www.youtube.com/watch?v=vAgEf4jX_1o