PSI’s David Turturean used early access to Anthropic’s Fable 5.2 to produce a solution to the Aaronson–Ambainis conjecture, an 18-year-old question about quantum speedups.
The proof was converted into a form checked by Lean.
Scott Aaronson believes it is correct, though he says he does not fully understand it.
The result rules out one possible route, not every route, to quantum advantage.
The solution seems to rule out one possible way in which quantum computers could solve some difficult problems faster. But it doesn't rule out every possible way.
Turturean is a recent MIT graduate and member of PSI's technical team. He put aside his plans to pursue graduate studies at MIT's computer science and artificial intelligence lab to work at the startup.
Scott Aaronson, one of the two researchers who asked the question, said he doesn't fully understand the proof, but he believes it's correct. The proof was also converted into a form that can be checked by a special program. PSI shared it with a group of leading researchers, who explained and studied it.
PSI develops research systems that combine artificial intelligence models with specialized tools, memory systems, and simulation infrastructure. Today, its revenue comes from using these systems to optimize data centers, taking into account heat, energy, water, and noise.
↳
READ ALSO
HACKS.GR / RELATED
AI Unlocked
OpenAI introduces Business Hub for business profiles and digital assistants
↗
Its long-term goal is to turn discoveries in theoretical physics into technologies that can be used in practice. PSI CEO Matt Pines said that theoretical physics has not yet become primarily a machine field, but he believes that the evolution is moving in that direction.
Artificial intelligence has already contributed to important solutions in mathematics and computer science, where results can be formulated and tested in code. Applying it to less abstract fields, such as physics and the life sciences, is more difficult. PSI focuses on how the results of calculations can be tested experimentally, as efficiently as possible.
The week before the article was published, startup Discovered Materials announced a $1 million challenge for AI-powered research into topics like superconductivity and heat dissipation.
PSI held the result until the Anthropic model was ready for release, then shared it with researchers working on the topic so they could better understand and test it.
The practices for the new speed of scientific research are still being shaped, said Jim Halverson, a professor at Northeastern University and a consultant to PSI. He was scheduled to attend a conference at MIT next week, where leading physicists would discuss how the scientific community can respond to the new pace of research.
>Stay connected with Hacks.gr
Set us as a preferred source on Google and find us on social media.
Preferred source
↗
end of article
Top comments (0)