NoticeThis site demonstrates one possible use of this domain. For acquisition, partnership, or investment inquiries, please use our contact form.

Future of Science

AGI in Physics and Mathematics

How AI proof assistants, simulation, and pattern discovery are augmenting research in physics and mathematics.

fig / physics & maths// field plate
Risograph illustration of human and machine scientific capability
Plate / Scientific discovery is moving into a new tempo across multiple fields.

Executive summary

Physics and mathematics are being augmented by AI in distinct ways. Proof assistants are reaching expert level on bounded problems. Simulation is being accelerated across high-energy physics, climate, and astronomy. Pattern discovery — particularly in large experimental datasets — is being reshaped.

Key concepts

  • AI proof assistants
  • Simulation acceleration
  • Pattern discovery
  • Theoretical frameworks
  • Experimental design

Proof assistants

AI proof assistants like AlphaProof have reached competitive performance on mathematical olympiad problems and contribute to active research.

Simulation

AI surrogates accelerate simulation in fluid dynamics, climate, astronomy, and high-energy physics. Some workflows are 100×+ faster with comparable accuracy.

Pattern discovery

Large experimental datasets in astronomy and particle physics are being mined more effectively by AI methods.

Theoretical work

AI is not yet producing major theoretical frameworks but is contributing to conjecture generation and exploration.

Key takeaways

  • 01AI proof assistants are at competitive expert level on bounded problems.
  • 02Simulation acceleration is producing 100×+ speedups in some workflows.
  • 03Pattern discovery in large datasets is improving.
  • 04Major theoretical advances remain primarily human.

Frequently asked questions

Could AI prove the Riemann hypothesis?

Possibly within the next decade or two. Several major mathematicians treat this as no longer fanciful.

What about new physics?

AI is contributing to data analysis and conjecture generation. Frameworks for new physics remain a deeply human pursuit so far.