Well, a general response to this sentiment is actually what the likes of Neil DeGrasse Tyson always say — you might not know now what the practical benefits of a theoretical discovery might be in the future, but chances are that more understanding of the universe will lead to practical technological developments.
This sort of objection was raised back when Faraday made a needle move remotely with electromagnetism, and now we have all of the electronics, surely beyond what even he could imagine. Something as arcane as quantum tunneling is what makes touchscreens on mobile devices work, and the theory of relativity is essential for the GPS tracking system.
More specifically, a simulated universe may allow specific technological exploits as well depending on how exactly it is being simulated. It would mean in a very real sense that there are specific additional dimensions to reality that we’re not able to directly experience, but which tie into the processes that we can experience and manipulate.
It can also serve as a tangible way in which science could work toward integrating its material quantitative side with its subjective qualitative side. In a simulation, meaningful phenomena are definitely no less physical than material ones, and probably would be connected together in some abstract mathematical way that would probably manifest as fractals. In a simulation, you have a solid framework for asking scientific “why” questions.
As for what I personally think could be studied and applied for practical purposes, there’s a lot going on in the universe that we currently think of as random. In a simulation, true randomness does not exist, at least not beyond the player interaction. Looking at the same processes as if they were governed by computer algorithms instead of looking at them as bottom-up manifestations of particle physics may uncover some hidden organizing mechanisms that impact which specific outcome happens when.
After all, probability only means that certain outcomes happen with a certain frequency. We use it when we cannot explain how exactly each individual outcome happens. Knowledge like this would definitely give us greater predictive powers, even if we couldn’t affect the algorithms themselves in any way. This is also where simulation hypothesis begins to get most easily testable, as you can hypothesize that if there’s a particular ordering mechanism in play, certain things would happen at certain times or in a certain order, in addition to particle physics doing their own thing.
On a personal level, it could get especially practical, since as I said in some earlier article, simulation can process “meaning physics” as well as material physics, separately or in a connected way. Just like there are laws that govern star formation or weather, there could be laws that govern subjective phenomena in a universal and mathematically describable way.
Funny thing is that “meaning physics” can simply be a real thing even if this world is emergent, but it would take a hypothesis with material roots like this one to get the material scientists and especially mathematicians to consider the possibility seriously and approach it in a way that can lead to its discovery. The surest way to not know anything (and gain no practical benefits) is to not explore possible hypotheses. Any judgments at this point are simply premature, though understandably skeptical.