What if the work these folks do to increase market efficiency does produce real-world value, as slower verifiable economic information would increase periods of uncertainty and price slippages?

Or the HFT engineers producing research and products that speed up general compute, such as accelerating SHA and ED25519 signature verification, so that their trading shops get information faster?

I wonder if there are actually a lot of third-order, downstream effects that society gets by paying engineers to just make computers and telecommunications faster.