AI helps prove a long-standing quantum computing conjecture

Solving a quantum mystery

Physical Superintelligence Inc. (PSI) announced on Friday that it successfully proved the Aaronson-Ambainis conjecture. This specific problem has remained an open question in the field of quantum computing since 2008. The conjecture is essentially an educated guess that researchers have been trying to verify for over a decade.

The research lab used a specialized system to tackle the problem, which included an unreleased model from Anthropic. While the specific model used by the lab has not been identified, the results were achieved in only three days. PSI described the achievement as settling one of the most fundamental questions in the field.

According to the lab, the proof clarifies that quantum computers have inherent limits. While these machines are still expected to be a major upgrade for certain tasks, the proof shows that specific exponential advantages in quantum computing require very particular patterns or structures. This moves the conversation from theoretical speculation to a confirmed mathematical reality. The lab noted that while quantum computing still has plenty of mysteries, there is now one fewer.

The role of AI in research

This breakthrough is part of a growing trend where artificial intelligence is being used to solve complex problems at the frontier of science. In September, OpenAI reported using AI to solve the Navier-Stokes equations, which is a 90-year-old math problem. OpenAI called this one of the deepest problems at the frontier of mathematics.

The use of AI in these high-level proofs is changing how researchers approach difficult questions. By integrating advanced models into research systems, labs can process complex variables and structures more quickly than traditional manual methods. This marks the latest breakthrough in mathematics achieved with the help of AI, leading many in the field to speculate about what this disruption will mean for the future of the discipline.

The speed of these results is particularly notable. PSI was able to finalize the proof in a three-day window, a timeframe that would likely be impossible for human researchers working alone. This suggests that AI is becoming a powerful accelerator for theoretical physics and advanced mathematics. It provides a way to navigate the vast search space of mathematical possibilities that human intuition alone might struggle to map out.

Impact on the math community

The rise of AI in solving deep mathematical problems is causing a shift in the academic landscape. Some prominent figures, such as UCLA professor Terence Tao, have noted that math's "romantic era" was over. He suggested that while exciting new things are possible, some things are being lost as the field changes.

This shift is already affecting the next generation of scientists. A study by MIT researchers found that 76% of graduate students and postdocs said they were unlikely to stay in academic math after they finish their studies. The same percentage of students reported that AI developments have made them feel less motivated to conduct research.

Students are increasingly reconsidering their career paths as AI takes on roles that previously required years of human dedication. As models become more capable of handling complex proofs, the traditional path of academic research is being re-evaluated. This creates a tension between the excitement of new technological capabilities and the preservation of traditional human-led mathematical exploration.

Expert reactions and context

Researchers who had previously worked on the problem described the proof as a huge result that resolved a decades-old question. According to PSI's announcement, a group of scholars from Stanford University described the proof as coming together in a rather magical way. This sentiment reflects the surprise and awe surrounding the speed and success of the AI-assisted proof.

While quantum computers exist today, they are still in the early stages of development. The proof provides a necessary roadmap for where the technology can and cannot go. By identifying the specific structures required for exponential advantages, the research helps prioritize which types of problems are best suited for quantum hardware versus traditional silicon-based systems.

A technical staff member at PSI built the research framework that utilized the unreleased Anthropic model to complete the proof. This collaboration between human-designed systems and advanced AI models highlights a new methodology for scientific discovery. It suggests a future where the most difficult questions in physics and math may be solved by systems that combine human goals with machine-scale processing.