An AI Just Solved an 18-Year-Old Quantum Physics Problem

A recent MIT graduate just used an early AI model to close the book on an 18-year-old quantum computing puzzle. David Turturean solved the Aaronson-Ambainis conjecture using early access to Anthropic’s Fable 5.2 model while working at the startup lab Physical Superintelligence. The problem concerns whether quantum computers can always outperform classical ones.

Scott Aaronson, one of the original researchers who posed the question, admits he cannot fully understand the machine-generated proof himself. He believes it works because the solution was autoformalized in Lean, a coding language for mathematics, and vetted by a team of human experts. This automated proof shuts down one theoretical path for high-speed quantum calculations, marking a major negative result for the field.

Our read: when machines can bypass human cognitive bottlenecks in pure math, theoretical physics risks transforming from a human endeavor into automated execution almost overnight. Physical Superintelligence currently funds its operations by using similar technical harnesses to optimize data center cooling and power infrastructure. The long-term goal, however, is turning purely theoretical breakthroughs into physical technologies.

Translating pure math into the physical world remains difficult because nature requires physical experimentation to validate models.

As startup labs race to commercialize machine-generated science, academic communities are rushing to establish peer-review standards for software-driven discoveries.

Watch for how upcoming academic summits redefine what constitutes a published scientific contribution when the author is effectively an algorithm.