Claude solves two open black hole physics problems

Original: Claude solves two open problems in the physics of black holes

Why This Matters

AI solving decades-old physics problems signals a new role for LLMs in research.

Anthropic's Claude Fable 5.2 solved two open theoretical physics problems involving black holes and quantum gravity, completing a calculation first set up in 2003 and finding an exact limit for a quantum gravity testing technique.

Claude Fable 5.2 has solved two open problems in theoretical physics, both connected to the search for a unified theory of quantum gravity. The findings, published as preprints in November 2026, emerged from Claude operating as a small research group of copies of itself.

The first problem centers on the Hawking-Page transition. In 1983, Stephen Hawking and Don Page showed that heating a model universe causes an abrupt phase transition—like water boiling—leading to black hole formation. Their calculation required general relativity, which breaks down at small scales where spacetime loses definite shape. In 1999, physicists found a phase transition in this regime using quantum methods, but couldn't determine whether it was abrupt or smooth. A 2003 group formulated the calculation needed to answer that question; it was never completed. Claude finished it, confirming the transition remains abrupt even when spacetime has no definite geometry.

The second problem involves the modular bootstrap, a technique used to test candidate quantum gravity theories. Claude determined the exact limit of its simplest version—a result that had remained unknown.

Holographic duality, the mathematical framework underlying both problems, allows physicists to translate hard gravitational questions into quantum mechanics problems, sometimes making them tractable. Claude's approach apparently leveraged that translation in both cases.

Source

anthropic.com — Read original →