My interpretation of the argument is not that it is equating atoms to $. Rather, it invokes whatever computations are necessary to produce (e.g. through simulations) an amount of value equal to today's global economy. Can these computations be facilitated by a single atom? If not, then we can't grow at the current rate for 8200 years.

Thanks for your detailed reply. Absolutely, there is some academic reward available from solving problems. Naively, the goal is to impress other academics (and thus get published, cited), and academics are more impressed when the work solves a problem. 

You seem to encourage problem-solving work, and point out that governments are starting to push academia in that direction. This is great, and to me, it raises the interesting question of optimal policy in rewarding research. That is supremely difficult, at least outside of the commercialisable. My understanding is that optimal policy would pay each researcher something like the marginal societal benefit of their work, summed globally and intertemporally forever. How on earth do you estimate that for the seminal New Keynesian model paper? Governments won't come close, and (I imagine) will tend to focus on projects whose benefits can be more easily measured or otherwise justified. So we are back to the problem of misaligned researcher incentives. But surely a government push towards impact is a step in the right direction.

Until our civilisation solves that optimal policy problem, I think academia will continue to incentivise the pursuit of knowledge at least partly for knowledge's sake. I wrote the post because understanding the implications of that has been useful to me.

I should clarify - I don't mean a small amount of work, but a small conceptual adjustment. The example I give in the post is to adjust from fully addressing a specific application to partially addressing a more general question. And to do so in a way that is hopefully intellectually stimulating to other researchers.

In my own work, using a consumer intertemporal optimisation model, I've tried to calculate the optimal amount for humanity to spend now on mitigating existential risk. That is the sort of problem-solving question I'm talking about. A couple of possible ways forward for me: include multiple countries and explore the interactions between x-risk mitigation and global public good provision; or use the setting of existential risk to learn more about a particular type of utility function which someone pointed me to for that purpose.