A Quantum Physicist Is Using OpenAI's o1 to Tackle Physics' Biggest Questions

Quantum physicist Mario Krenn is using OpenAI's o1 model to help answer what the source describes as life's biggest questions — a concrete signal that frontier reasoning models are now being tested by working scientists inside one of physics' hardest disciplines, not just by hobbyists and developers.

The detail comes from a short item titled "Answering quantum physics questions with OpenAI o1," which states plainly that Krenn, a quantum physicist, uses OpenAI o1 to help answer these questions. The item is sparse. It names one person, one model, and one domain. But the combination is what makes it worth attention: a professional researcher in quantum physics — a field where correctness is unforgiving and mathematics is dense — is treating a general-purpose AI model as a working assistant on deep problems.

Who is doing this, and with what

The facts, as reported: Mario Krenn is a quantum physicist. The tool he is using is OpenAI o1. The purpose is to help answer big questions — the source frames them as life's biggest, and in the context of Krenn's profession, those questions sit in quantum physics.

OpenAI released o1 as its first model family optimized for reasoning — models designed to spend more compute thinking through a problem step by step before producing an answer, rather than responding immediately. That design choice is precisely why a physicist would reach for o1 rather than a conventional chatbot. Quantum physics problems tend to reward deliberate, structured reasoning: derivations that stretch across many steps, constraints that interact, and answers where a single logical slip invalidates the whole chain.

The source does not specify which questions Krenn is putting to the model, what results he has obtained, or whether the outputs have made their way into published research. What it establishes is the act itself: a domain expert at the level of quantum physics is incorporating o1 into his process of grappling with fundamental questions.

Why a two-sentence story matters

The significance lies in the pattern it represents. For most of the last decade, AI systems have been demonstrably useful at writing boilerplate, summarizing documents, and drafting code — tasks where errors are cheap to catch and consequences are low. Fundamental physics is the opposite case. Errors are subtle, expertise is scarce, and the questions themselves are the kind that define research programs and careers.

When a working quantum physicist publicly uses a model like o1 on such questions, it marks a shift in where AI assistants are being pointed. The user is no longer a generalist asking for an explanation of entanglement for a blog post. He is a specialist using the model as a collaborator on the specialist's own terrain, where he can judge the quality of the output with expert eyes.

That last point matters more than it might seem. The most credible evaluations of AI capability in science do not come from benchmark leaderboards alone. They come from practitioners who can tell the difference between a fluent answer and a correct one. A quantum physicist applying o1 to quantum physics questions is, whether framed that way or not, running exactly that kind of expert evaluation — with his own research questions as the test set.

The stakes for AI in science

The research stakes here are straightforward. Reasoning models such as o1 represent the current frontier of the industry's effort to make AI systems useful on problems that cannot be solved by pattern-matching against training data. Mathematics, physics, and other formal disciplines are the natural proving ground for that effort, because success and failure are unambiguous. An answer to a physics problem is either right or wrong, and the field's practitioners are unusually well equipped to tell the difference.

If models like o1 can genuinely contribute to how physicists explore hard problems — even as assistants that propose angles, check reasoning, or surface approaches a human might not have tried — the implications extend beyond any single research group. Science has always been bottlenecked by expert time. A tool that multiplies what one expert can examine shifts that bottleneck.

The market and policy context reinforces the attention. OpenAI has positioned o1 as evidence that a new scaling direction — more compute at inference time, spent on thinking rather than on training alone — can keep delivering capability gains. Every credible report of a domain expert deriving value from these models is a data point in that argument, and quantum physics, with its reputational hardness, is a louder data point than most.

At the same time, one physicist's use of a tool is an anecdote, not a study. The source reports the activity, not validated results. Whether o1's contributions to Krenn's work prove substantive or merely convenient remains, on the basis of this item, an open question — one that only further reporting and Krenn's own assessments can answer.

What to watch

The trajectory to watch is whether this pattern generalizes. The item describes one quantum physicist using one model. The interesting future is the one in which this becomes ordinary: physicists, mathematicians, and other formal-science researchers routinely routing hard problems through reasoning models, and reporting honestly on where the models help and where they hallucinate.

For OpenAI, cases like Krenn's are early evidence for o1's core pitch — that reasoning-optimized models can hold their own on problems that defeated earlier generations of AI assistants. For the research community, they are an invitation to test that claim rigorously. And for anyone tracking where AI actually earns its place in serious work, the name to follow next is Mario Krenn, and the question is what his biggest questions look like after a physicist and a reasoning model have spent real time on them together.