Sakana AI Hires Jürgen Schmidhuber as Chief Scientist

Sakana AI has hired Jürgen Schmidhuber as Chief Scientific Advisor, giving the Tokyo startup direct access to the researcher it calls the "father of modern AI." Schmidhuber's resume stretches back decades: a 1987 thesis on meta-learning and recursive self-improvement, early work on world models in 1990, and the 1991 deep learning techniques that underpin much of the current AI industry. Sakana says its own projects, including the Darwin Gödel Machine and The AI Scientist, trace directly back to that body of work.

The hire lands as Sakana stands up a new RSI lab, short for recursive self-improvement, built around the idea of a research loop where machine intelligence keeps refining itself without constant human input. Sakana wants to pull together what it calls a "critical mass of world-class experts" in Tokyo to run it, and Schmidhuber's involvement is meant to anchor that effort. The framing from Sakana leans heavily on Japan's history in neural network research and robotics, positioning the company as a bridge between that legacy and whatever comes next in physical AI. Schmidhuber laid out his own read on that mission in the following statement.