Mar 13, 2026

Solving the Wrong Problem Works Better - Robert Lange

Robert Lange from Sakana AI discusses Shinka Evolve, a framework combining LLMs with evolutionary algorithms for open-ended program search. The conversation explores how Shinka Evolve addresses the limitations of systems like AlphaEvolve by co-evolving problems and solutions, its sample-efficient architecture using UCB bandits and quality-diversity search, and its applications in circle packing, competitive programming, and evolving MoE loss functions. The discussion also delves into the philosophical debate on whether these systems produce true novelty or are parasitic on their starting conditions, and the future role of the "AI Scientist" as a human co-pilot.

Nov 23, 2025

AI That Can Change Its Mind? (New Architecture) [w/ Sakana CTO]

Llion Jones, a co-inventor of the Transformer, and his colleague Luke Darlow from Sakana AI argue that the AI industry is trapped in a local minimum by the Transformer's success. They discuss the architecture's fundamental limitations using the 'spiral problem' analogy and introduce their new, biology-inspired Continuous Thought Machine (CTM), an architecture designed for more human-like, sequential reasoning and adaptive computation.