On a Monday in late October, OpenAI announced that its artificial intelligence models had produced three discoveries spanning harmonic analysis, complexity theory and quantum field theory, a claim that arrives in a long human tradition of mathematicians and physicists testing their ideas against the slow judgment of their peers. The headline result concerns the local smoothing conjecture, a question about how waves and functions can be decomposed into simpler parts, which the company says it has now settled in every spatial dimension, where it was previously understood only in three. Whether t
OpenAI Says Its Models Made Three Scientific Discoveries in Mathematics, Computer Science and Physics
A twinge of sadness, then a realisation of how much can now be accomplished
Mark
Three fields in one announcement. What's the piece most people will point to?
Mimi
The local smoothing conjecture. OpenAI says its models proved it in every spatial dimension. Before, it was only proved in three.
Mark
And someone close to the field has weighed in?
Mimi
Alex Iosevich at Rochester. The conjecture came from his own PhD adviser, Chris Sogge. He called it a historic development and said it was a shock at first.
Luke
Worth noting that's one expert's reaction, and he's reacting to OpenAI's claim. The reporting doesn't say the proof has been independently checked.
Mark
He also mentioned sadness, didn't he?
Mimi
A twinge, yes. He said many theoretical mathematicians feel it, but he expects it to give way to a sense of what's now possible.
Mark
What about the physics claim?
Mimi
OpenAI says it provided rigorous proof of the existence of quantum chromodynamics. The physics group is delaying release so it can write its own interpretation.
Luke
That means the public can't yet examine the physics data. And the claim is OpenAI's wording; we don't have an outside assessment of it here.
Mark
Why use those three expert groups, the shepherds?
Mimi
OpenAI says it wanted experts to contextualise results. It also admits that picking them shapes who gets involved first and creates an information asymmetry.
Luke
And the Anthropic enzyme claim last month drew mixed reactions, with some saying it was exaggerated. So the pattern is a caution, though no one has said the same about this one yet.
Il Polso
- OpenAI has declared three AI-driven breakthroughs, but no independent expert has yet confirmed them, leaving a tense gap between bold claim and proof.
- A mathematician who studies the field called the result historic, while his own adviser's conjecture sits at the center of the news, stirring both pride and unease in the research community.
- The company is claiming it disproved the Exponential Time Hypothesis and gave a rigorous proof of quantum chromodynamics, two results that would reshape computer science and particle physics if they hold.
- One physics group has chosen to pause publication to write its own interpretation, while other groups released their raw data at once, revealing a split over how AI findings should be shared.
- Skepticism lingers from a recent claim by Anthropic about an enzyme, and the verdict will rest on peer review that has not yet been completed.
On a Monday in late October, OpenAI announced that its artificial intelligence models had produced three discoveries spanning harmonic analysis, complexity theory and quantum field theory, a claim that arrives in a long human tradition of mathematicians and physicists testing their ideas against the slow judgment of their peers. The headline result concerns the local smoothing conjecture, a question about how waves and functions can be decomposed into simpler parts, which the company says it has now settled in every spatial dimension, where it was previously understood only in three. Whether these machines have truly added to human knowledge, or have produced results that still need the patient scrutiny of experts, remains an open question that the coming months will answer.
On Monday, OpenAI announced that its AI models had made three scientific discoveries, in harmonic analysis, complexity theory and quantum field theory. The San Francisco company, valued at $852bn, framed the results as advances in mathematics, computer science and physics, and independent experts are still weighing them.
The headline claim concerns the local smoothing conjecture, a central idea in harmonic analysis about how complex signals can be broken into elementary waves. OpenAI says its models proved it in every spatial dimension, where researchers had only managed three. Alex Iosevich, a mathematics professor at the University of Rochester, told Al Jazeera that the company had appeared to "crack" the problem, calling it a "historic development." The conjecture was formulated by his own PhD adviser, Chris Sogge, which gives the news a personal weight. At first it shocked him, he said, but he now feels joy and optimism about what AI might do for science. He also admitted a twinge of sadness, one he believes many theoretical mathematicians share, and expects the field to change drastically.
OpenAI also claims to have disproved the Exponential Time Hypothesis and to have produced a rigorous proof of the existence of quantum chromodynamics. To release the work, the company consulted outside experts it called "shepherds." The mathematics and computer science groups published their AI-generated data immediately, while the physics group is holding back to write its own interpretation first. OpenAI acknowledged the tradeoffs, noting that outside collaboration brings expertise but also shapes who sees the results early.
The announcement follows Anthropic's claim last month of a CRISPR-like enzyme, which drew mixed reactions. Whether OpenAI's findings endure will depend on verification and peer review, which have not yet been completed.
On Monday, OpenAI announced that its AI models had made three scientific discoveries, spanning harmonic analysis, complexity theory and quantum field theory. The San Francisco company, valued at $852bn, presented the results as a set of findings in mathematics, computer science and physics. Independent experts are still evaluating them.
The headline claim concerns the local smoothing conjecture, one of the central propositions in harmonic analysis. That branch of mathematics studies how complex signals and functions can be broken down into elementary waves. According to OpenAI, its models proved the conjecture in every spatial dimension for the first time. Researchers had previously established it only in three dimensions.
Alex Iosevich, a mathematics professor at the University of Rochester who specialises in the field, told Al Jazeera that OpenAI appeared to have "cracked" the problem, and he called it a "historic development." He has a personal stake: the conjecture was formulated by his PhD adviser, Chris Sogge. The news, he said, was a shock at first, and he now feels joy and optimism, expecting science as a whole to make giant leaps because of AI.
Iosevich also expects theoretical mathematics to "change drastically." He admitted to a twinge of sadness, one he said many theoretical mathematicians he knows share. He believes it will give way to a realisation of how much can now be accomplished. Problems in pure mathematics, he noted, often grew out of concrete physical questions, and those connections can now be explored quickly and decisively.
OpenAI made two further claims. Its models disproved the Exponential Time Hypothesis, which concerns how long it takes to solve complex computational problems. They also provided what the company called a mathematically rigorous proof of the existence of quantum chromodynamics, the theory of how quarks and gluons interact.
The company said it reached out to scientists outside OpenAI to turn its raw research data into scientific understanding. It chose different release methods after consulting three expert groups it designated "shepherds." The groups working on the mathematics and computer science problems opted to publish the AI-produced data immediately. The physics group chose to hold back its release so it could first write its own interpretation of the findings.
OpenAI acknowledged that the approach has costs. Its statement said all such approaches involve tradeoffs: collaborations let experts unpack the results and place them in the wider academic community, but choosing the first shepherds shapes who gets involved early and creates an asymmetry of information. The company said it welcomes feedback on how to broaden participation.
The announcement follows other claimed AI breakthroughs. Last month, Anthropic, maker of Claude, announced the discovery of a new enzyme said to resemble CRISPR, the gene-editing mechanism used in treating conditions such as cancer and sickle cell disease. Scientists reacted unevenly: some hailed it as exciting, while others said the company had overstated its importance. Whether OpenAI's results fare better will depend on independent verification and peer review, which have yet to be completed.
Citazioni salienti
The conjecture was formulated by my PhD adviser, Chris Sogge, so it is near and dear to my heart, and the news was a shock at first.— Alex Iosevich, University of Rochester mathematics professor
All of these approaches involve tradeoffs.— OpenAI
Vuoi la storia completa?
Leggi l'originale su Al Jazeera ↗
War & Conflict