iCreatAI News
Back to main site
Skip to content
Only Industry#Research
Nov 11Wed
  1. TechCrunch · AI78

    PSI is trying to make a business out of AI-driven scientific discovery

    Physical Superintelligence (PSI) used early access to Anthropic's Fable 5.2 model to solve the eighteen-year-old Aaronson-Ambainis conjecture in quantum query complexity, a result Scott Aaronson said he believes is correct because it was autoformalized in Lean and reviewed by top researchers. PSI, which currently earns revenue by optimizing data centers, aims to convert theoretical physics breakthroughs into technologies, though CEO Matt Pines noted that physics differs from math because "nature gets a vote in your models of reality." The company held its result until Anthropic's model was nearly ready for release, sharing it with relevant researchers, as adviser Jim Halverson said best practices for such work are still emerging.

    Why it matters: It offers a concrete case of an AI-assisted proof in quantum complexity and a look at how a startup is trying to commercialize AI-driven discovery.

Nov 6Fri
  1. TechCrunch · AI58

    Axiom Math wants to keep AI slop out of mathematics

    Axiom Math, a $1.6B startup founded in March 2025 by Stanford math Ph.D. dropout Carina Hong, announced it has formally verified the Stacks Project, translating its roughly 7,654 pages into about 16 million lines of computer-checked Lean code using a fleet of AI agents. The company argues that verified results become building blocks for future proofs and that the approach can extend to software written by AI agents.

Nov 3Tue
  1. MIT News AI38

    Regina Barzilay and Jim Collins win Pritzker Prize for AI and Science Research Excellence for work in AI-enabled antibiotic discovery

    MIT's Regina Barzilay and James Collins received the 2026 Pritzker Prize for AI in Science Research Excellence for their joint work in AI-enabled antibiotic discovery. Their 2020 Cell paper with first author Jonathan Stokes described halicin, the first AI-discovered antibiotic, identified by a deep learning model trained on 2,500 molecules and shown to kill drug-resistant bacteria such as C. difficile, A. baumannii, and M. tuberculosis, with the discovery process completed in weeks rather than years. They later used machine learning to identify abaucin in 2023, a narrow-spectrum antibiotic effective against A. baumannii.

Oct 28Wed
  1. The Verge · AI71

    Researchers are getting tired of OpenAI gatekeeping its breakthroughs

    OpenAI announced new AI-generated results in mathematics (a claimed resolution of the Kakeya conjecture in all dimensions), computer science (a claimed disproof of the exponential time hypothesis), and physics (work on quantum chromodynamics and Yang-Mills theory), each with a different release strategy. Researchers told The Verge the results look significant but criticized OpenAI for handpicking expert "shepherds" and withholding the physics materials, with the Advisory Group on Mathematics and Artificial Intelligence urging immediate public release of the natural-language and Lean outputs.

  2. TechCrunch · AI88

    Frontier AI labs say their models refuted the Exponential Time Hypothesis. What does that mean?

    OpenAI released solutions to three major math problems, including a claimed refutation of the Exponential Time Hypothesis (ETH), and Anthropic later released its own refutation of ETH. MIT professor Ryan Williams, working with ETH originators Russell Impagliazzo and Ramamohan Paturi, is organizing the computer science community's response, releasing an open letter calling for feedback and discussion of norms around AI contributions. Williams said OpenAI checked whether the result breaks any cryptographic systems and was reasonably confident it does not.

    Why it matters: Explains what a claimed AI refutation of a foundational complexity assumption means and how researchers are responding.

Oct 27Tue
  1. The Decoder88

    OpenAI asks outside experts, including a Nobel laureate, to check proofs its AI produced in math, computer science, and physics

    OpenAI has asked outside experts to act as "shepherds" for AI-generated results in three fields, including a claimed disproof of the Exponential Time Hypothesis, a proof of the local smoothing conjecture in every dimension, and a proof that certain quantum field theories exist in four dimensions. The math and computer science proofs were translated into Lean, though Lean only checks that a proof establishes the statement as written, not that the statement captures the intended problem. A group led by Nobel laureate David Gross will shepherd the physics work, while the ETH result is being independently checked by Russell Impagliazzo, Ramamohan Paturi, and Ryan Williams, who have no NDAs, are unpaid, and say their involvement is not an endorsement.

    Why it matters: It offers a detailed look at how OpenAI is handling extraordinary AI-generated claims, including the limits of Lean verification and the open questions about credit and access raised by the reviewers themselves.

  2. The Verge · AI88

    OpenAI announces three big science discoveries and a new way to work with academics

    OpenAI released three claimed "major" results — a proof of Sogge's local smoothing for the Euclidean wave equation (also yielding the Kakeya maximal and dimension conjectures and Stein's spherical restriction estimates), a deterministic subexponential-time algorithm for fixed-width CNF counting that would refute the Exponential Time Hypothesis, and a mathematically rigorous existence proof of certain interacting quantum field theories in four dimensions including QCD. External experts will shepherd the computer science and physics results, while the math output is released directly for researchers to work through; OpenAI also acknowledged an AGMAI panel's recommendation that AI companies not adopt the practice of designating certain people to receive output first.

    Why it matters: It documents OpenAI's latest claimed breakthroughs and the emerging norms around expert review and early access for AI-generated results.