In a remarkable demonstration of artificial intelligence's growing capabilities, a University of Pennsylvania statistics professor has reportedly used OpenAI's GPT-5.6 Sol Pro to disprove a long-standing conjecture in statistical theory—something that had stumped human researchers for over three decades.
The breakthrough centers on the Benjamini-Hochberg method, a widely used technique for controlling false discovery rates in multiple hypothesis testing. For 30 years, researchers have been trying to resolve a key conjecture related to its behavior under certain conditions, but without success. The new AI model, however, reportedly found a solution in just 90 minutes—while its predecessor, GPT-5.5, had failed to produce any progress even after 20 hours of intense computation.
AI Reconstructs Knowledge in Novel Ways
The solution, according to the professor, was not a simple recitation of known facts, but rather a creative synthesis of existing methods applied in a previously unexplored manner. This raises fascinating questions about the nature of AI-generated insights: is this a true leap in understanding, or simply a sophisticated recombination of known elements?
The feat underscores a pivotal shift in AI's role within scientific research. While earlier models were primarily used for assistance and idea generation, GPT-5.6 Sol Pro appears to be capable of independently arriving at significant theoretical breakthroughs. The implications are profound, especially as AI systems become more adept at navigating complex mathematical and statistical domains.
What Comes Next?
Although the specific details of the solution have not yet been published, the incident has sparked a broader conversation about AI's potential to contribute to fundamental scientific discoveries. As AI systems continue to evolve, they may increasingly serve not just as tools for researchers, but as collaborators in the advancement of human knowledge.
Whether this is the beginning of a new era of AI-assisted science or a one-off triumph remains to be seen. But for now, it stands as a powerful testament to the rapid progress in AI capabilities—and the evolving boundaries of what machines can achieve.
