Episode 11: When AI Starts Improving Itself

Описание: In Episode 11 of Anthony and Greg on Intelligence (AGI), Anthony Moggio and Greg Prost tackle one of the most important shifts happening in AI right now: models that are beginning to improve themselves — faster than humans ever could.

They break down what’s changed in the latest releases from Google Gemini and OpenAI (ChatGPT 5.1 / 5.2), and why these updates feel fundamentally different. From AI taking longer “thinking time,” to running deep research for nearly an hour, to discovering inefficiencies in its own logic and energy use, this episode explains what’s really happening behind the scenes.

Anthony and Greg also discuss how AI is rewriting parts of its own code, drastically reducing compute and power requirements, and accelerating at a pace that’s hard to slow down. They explore what this means for data centers, competition between AI companies, and why we may already be watching the start of an exponential snowball effect.

The episode wraps with a thoughtful conversation about AI throttling, intelligence ceilings, and why today’s cutting-edge tools may already be outdated the moment you log in.

If you want to understand where AI is headed — not years from now, but right now — Episode 11 is essential listening.

Subscribe for grounded, real-world conversations about AI, technology, and what comes next.