Artificial intelligence just helped solve one of the biggest unsolved problems in quantum computing.
Physical Super Intelligence Inc. (PSI) said on Friday It uses Anthropic’s unpublished model to prove the Aaronson-Ambainis conjecture, a long-standing problem in quantum computing.
The physics research lab said the conjecture, essentially an educated guess that has yet to be confirmed, has been open since 2008 and involves determining when quantum computers could become much faster than ordinary computers.
The laboratory developed a research system that included an anthropic model. PSI said the system was able to prove the conjecture in three days, “solving one of the most fundamental problems in quantum computing.”
“There are still many unanswered questions about quantum computing, but one is now gone,” the lab said.
The announcement marks the latest breakthrough in artificial intelligence in mathematics. In September, OpenAI said it had solved a 90-year-old mathematical problem, the Navier-Stokes equation, calling it “one of the most profound problems at the forefront of mathematics.”
Artificial intelligence’s disruption of mathematics has many in the field speculating about what it means for their future. Terence Teresa, a distinguished mathematics professor at UCLA, said last month that the “romantic era” of mathematics was over, adding, “We are entering an era where exciting new things are possible, but something is disappearing.”
Mathematics students are also rethinking their careers. Researchers at MIT polled graduate students and postdocs and found that 76% said they were less likely to pursue academic mathematics jobs after graduation, while the same number said advances in artificial intelligence have made them less motivated to pursue research.
PSI technical staff member David Turturean developed the research system to prove the Aaronson-Ambainis conjecture using an unpublished anthropic model.
It’s unclear what unreleased model PSI is using. Anthropic did not respond to a request for comment.
PSI said evidence suggests that some of the exponential advantages of quantum computing require special patterns or structures, implying that quantum computers have limitations that have long been suspected (although they remain a significant upgrade for some tasks). Although quantum computers already exist, they are still in the early stages of development.
Researchers who have studied the issue before said the evidence was a “huge result” that solved a decades-old problem.
“It all came together in a pretty magical way,” Stanford’s Guy Blanc, Jordan Docter, Carmen Strassle and Li-Yang Tan said of the proof, according to a statement from PSI.