Sakana AI brings on Jürgen Schmidhuber as Chief Scientific Advisor for new RSI lab
Tokyo-based startup Sakana AI has appointed Jürgen Schmidhuber as Chief Scientific Advisor. He will lead a new RSI lab focused on self-improving research systems for machines.
Tokyo-based startup Sakana AI announced that Jürgen Schmidhuber is joining in the role of Chief Scientific Advisor. According to Sakana, he is a researcher whom the company itself refers to as the "father of modern AI" – crediting him with early work on world models from 1990 and deep learning techniques from 1991, on which much of today's AI boom is built. His 1987 dissertation on meta-learning and recursive self-improvement, according to Sakana, kicked off nearly four decades of research into how machines can learn to learn.
According to the company, Schmidhuber will help lead the new RSI lab at Sakana, whose goal is to build a self-improving research cycle through which machines would progressively become smarter. Sakana states that it wants to assemble a "critical mass of world-class experts" in Tokyo. According to Sakana, his ideas have directly influenced the company's existing projects, specifically Darwin Gödel Machine and The AI Scientist.
Schmidhuber himself stated that Japan is the cradle of foundational neural network architectures and advanced robotics, and that the future of intelligence lies not only in language but in physical AI powered by world models. Sources do not provide details on the specific scope of the role, the timeline for starting, or financial terms.
Why it matters
This is a personnel strengthening of the research direction at Sakana AI, signaling that the company is focusing on self-improving AI systems (RSI) and physical AI/world models as an alternative to purely language-based models. For companies and investors tracking trends in AI research, it is an indicator of where the next wave of development after LLMs may be heading.
Relevant practical impact
What this means
For a business
Companies tracking the direction of AI research may view this talent acquisition as a signal of growing interest in self-improving research systems and world models as a further research direction alongside language models.
DevelopmentCheck the original
Event sources
only one source so far · 1 publisher, 1 independent. We count feeds from the same owner only once.