There is a reasonable argument that your question is at least NP, and plausibly NP-hard or harder depending on how you formalize the verification oracle.
Ask HN: What hard problems are still underexplored?
21–30 of 31 posts
Re: Ask HN: What hard problems are still underexplored?
#22Chemistry (Or biology, as an extension of it) simulations. Current tools include Newtonian atom-centered force fields that are fit to a specific situation and lose validity outside it, and quantum computations that are very slow, and don't scale well. I have a hunch there is something about the underlying physics we are missing, and that we have not hit the endgame of modelling physics at this scale.
I've been experimenting in this space, where might I find a guide for what to build that would be useful to you? I suspect most existing approaches are an order of magnitude slower and harder to use than they need to be.
Re: Ask HN: What hard problems are still underexplored?
#231) Designing built environments that maximize the community and enjoyment of the people who live in them
and perhaps even moreso 2) Figuring out how to get them built
It seems we mostly know the answers for 1, we just don't know how to get them built in a sea of development regulations and entrenched interests etc.
Re: Ask HN: What hard problems are still underexplored?
#24Perpetual-ish motion machines. While a true perpetual motion machine physically cannot exist, a machine that operates at an efficiency rate to be for all intents and purposes "perpetual" is theoretically possible, if not physical
Perpetual-ish motion machines exist if you know where to look.
Re: Ask HN: What hard problems are still underexplored?
#25Teleportation
Solved problem. https://x.com/AshtonForbes
Re: Ask HN: What hard problems are still underexplored?
#26Re: Ask HN: What hard problems are still underexplored?
#27Something trivial as matching people based on interest. We have social networks and various marketplaces but we are still unable to pair/match people based on desired activities, items for sale, services, relationships(online dating), even jobs.
Imagine you have some ancient toaster you are about to throw in the bin because it is old and you have no use for it and it has no value on the market. Yet, on the other side of the planet, there is a guy who is desperately looking for exactly this toaster because of #reasons. Yet, these people will never be able to find each other to trade.
Yes, there is ebay and whatnot, just like there is tinder for dating, facebook for socialising, various platforms for job hunting, but all these platforms are extremely inefficient in actually delivering on the promise of matching people based on the supply and demand.
The search engines all these platforms use are all very primitive and completely unable to provide the desired service. They are essentially all the same, they just cater to different markets. But there is little technical distinction among them.
The toaster example is a completely trivial one. You can easily expand it to a job where you need a person with specific skills and experience. But you will simply never be able to find that person via any of the existing pathways. Except sheer luck and word of mouth.
This can be likely solved via something like brain implants where we can be connected to the internet and immediately provide necessary context or answer some questions to build a better profile as a "supplier" or "buyer" that could allow a better match. But we're infinitely far away from it.
And this is just one of millions of such small problems that are really hard to solve.
The advertising companies all use tracking to try and mitigate this as much as possible so they can offer you the most likely product or a service that you actually might be interested in buying. But again, these are very primitive solutions.
Re: Ask HN: What hard problems are still underexplored?
#28Observability that can produce causal explanations rather than just timelines. We have great tooling for logs/metrics/traces, but very little that helps engineers understand why a distributed system behaved the way it did. Automated causal graphs for incidents still feel like an open problem.
That you know of.
Downvoting when someone gives information you didn't have makes them not want to give more information.
A word to the unwise is insufficient. https://www.paulgraham.com/word.html
Re: Ask HN: What hard problems are still underexplored?
#29Convincing 80% of developers to consider "alternative" solutions to the cargo cult seems like a pretty intractable task.
Re: Ask HN: What hard problems are still underexplored?
#30Chemistry (Or biology, as an extension of it) simulations. Current tools include Newtonian atom-centered force fields that are fit to a specific situation and lose validity outside it, and quantum computations that are very slow, and don't scale well. I have a hunch there is something about the underlying physics we are missing, and that we have not hit the endgame of modelling physics at this scale.
I've been experimenting in this space, where might I find a guide for what to build that would be useful to you? I suspect most existing approaches are an order of magnitude slower and harder to use than they need to be.
I would reject the premise that the field of molecular/biological simulation is underexplored nor that existing approaches are "slower and harder to use than they need to be". This is a field that has been explored, in fact, by the most brilliant minds and the difficulty arises more in theoretical considerations (that is, devising algorithms to faithfully approximate the developed physics) rather than an obvious no-brainer application of AI.
The field of molecular and biological simulation is far more than simply "Newtonian mechanics". There is indeed a field called molecular dynamics (MD) that relies on "classical mechanics" yet it's defined usually in the Lagrangian formalism. Furthermore, there has been tons of work over the past few decades in developing more accurate numerical approximation algorithms. There is a ton of a theory in this field and if you're interested, the "MD Bible" is "Understanding Molecular Simulation" by Daan Frankel.
Now, MD is just the tip of the iceberg. Almost all chemistry simulations are built entirely from making subtle approximations to quantum mechanics and carefully building up frameworks. For example, Hartree-Fock theory (HF), Density Functional Theory (DFT), Couple Cluster theory (CCSD(T)), etc. Then there is a field known colloquially as semi-empirical methods which are a sort of combination of the above two methods. And that's just on the side of chemical simulations (i.e. I'm excluding physics-specific simulations etc).
And now, more recently there has been effort in building machine-learned interatomic potentials, machine-learned density functionals, equivariant graph neural networks, etc etc.
If you're still interested in these class of problems, consider trying to build a good model for OMol25: https://arxiv.org/abs/2505.08762