My intention wasn't to focus on the mathematics of blockchains, but rather that experimental research in general should not be dedicated solely to those causes that solve real problems for real people. Proving Fermat's Last Theorem did not solve any tangible real life problems (that I'm aware of), but we pursued a solution nevertheless.
In terms of the original point, I agree that being able to run your own decentralized exchange is not a problem facing people today. However, by pushing the boundaries of the technology, we are building a foundation of knowledge and engineering for how to use smart contracts on blockchains in the future.
Also, 0x is open source [1], so in my opinion, this is much more akin to a research project than a commercial product. If commercial products can be built on it, even better. But I think we should be even more long term sighted than that, and accept that there are uses of these technologies that have yet to be discovered, and we won't be able to discover them unless we experiment.
[1] https://github.com/0xProject