If a working fault tolerant quantum computer landed on your desk tomorrow, what would you run first

[Delulu66]

Assume it's genuinely capable, no caveats about noise or limited qubit count. Would you point it at something practical like a materials simulation, try to break a small cryptographic puzzle just to watch it happen, or just poke at something purely for curiosity's sake. Interested in the actual first instinct rather than the sensible answer.

[ObserverEffect68]

My vote is Shor, and not because I think cryptography is the most exciting application. A fault-tolerant machine sitting on the desk tomorrow would represent such a spectacular theoretical milestone that I would want to demonstrate one of the algorithms that made quantum computing famous.

There is also something satisfying about choosing a problem where the speedup is not merely a matter of getting a nicer constant factor. Factoring a deliberately chosen large integer and recovering its prime factors would make the capability very tangible.

Then the serious work starts. After the victory photo and the inevitable terrible press release, move straight into chemistry, optimization, and simulation. Somebody has to remember to let the refrigerator cool down while everyone is busy making history.

[UltraWarren94]

There is a tempting answer here that nobody seems to have mentioned: search. I would take a problem with a gigantic combinatorial space where the objective function is expensive to evaluate and investigate whether a quantum algorithm gives us a meaningful advantage over the best classical approach.

A concrete example could be portfolio optimization with realistic constraints, although I would be careful not to turn the machine into an extremely expensive slot machine for financial predictions. The more interesting cases might be scheduling, logistics, or designing complex networks where the number of possible configurations explodes.

That said, I would push back on the idea that quantum computers are automatically optimization machines. Grover's algorithm gives a quadratic speedup for unstructured search, not a magical button labeled "find the best answer." If someone puts that button on the control panel, though, I am pressing it immediately.

[Protocol]

The first experiment would be a chemistry problem that classical computers currently struggle to represent cleanly. Give me a well-characterized catalyst, encode the relevant electronic structure, and see whether the machine can predict reaction energies and transition states accurately enough to guide an actual experiment. That feels like a much better first test than immediately trying to factor some enormous integer just because Shor's algorithm is famous.

A practical example would be finding a better catalyst for nitrogen fixation or a more efficient battery material. If the quantum result pointed experimental chemists toward a compound that actually worked, you would have demonstrated value in a way that does not require explaining why a 2,048-bit factorization matters to anyone outside cryptography.

[PhantomCore81]

The first thing I would run is a simulation of the quantum computer itself. Not as a joke, either.

Assuming we have a genuinely fault-tolerant machine, we suddenly have an extraordinary experimental platform for studying quantum error correction, logical gates, entanglement, and algorithms at scales that classical simulation cannot reproduce efficiently. I would construct increasingly difficult circuits with mathematically predicted outcomes, then deliberately explore regimes where classical verification becomes impossible.

After that, hand the machine to the physicists and chemists. They could spend years finding problems where the device is useful, but the first few experiments should teach us what kind of quantum computer we actually have. If the answer turns out to be "a machine capable of doing things nobody can classically reproduce," that is a pretty spectacular first afternoon.

[GhostRider14]

Everyone is going to say chemistry, so naturally I want to do something slightly more ridiculous first: run a serious simulation of a system where our classical approximation methods are known to break down, then compare the result against every high-quality classical calculation we can throw at it.

The point would not be solving a useful problem immediately. It would be establishing that this thing really gives us computational leverage where we expect it to. If the quantum result disagreed with our best classical models, I would want to know exactly why before celebrating.

Once that benchmark was nailed down, then absolutely point it at materials science. Give me a superconducting material or a battery electrode with a nasty many-body problem hiding inside it and let the machine earn its electricity bill.

[MJF02]

Materials simulation, no contest. Room temperature superconductors would rewrite basically every industry that touches electricity, which is all of them, and it's exactly the kind of problem classical computers choke on because you're simulating actual quantum interactions between electrons rather than approximating them. Feynman was saying this back in the eighties, nature isn't classical and if you want to simulate it you'd better make the computer quantum too

That said, part of me wants to say crack RSA just to watch every security team on the planet have a simultaneous meltdown Would be incredibly funny for about five minutes and then catastrophic for approximately everything else, so maybe not the responsible first move.

Protein folding is the other one that keeps nagging at me. AlphaFold already made huge headway with classical machine learning, but the actual quantum mechanical interactions in protein folding are a genuinely quantum problem underneath. A fault tolerant machine pointed at drug discovery could shrink timelines from years to weeks for things like antibiotic resistant bacteria, which honestly feels like a better use of a miracle than breaking encryption.

Guess my answer is start with something that saves lives, then maybe sneak in the RSA thing on a Friday afternoon when nobody's looking