[eigenket]

I agree with you that the general optimism towards quantum computing needs some tempering. It is certainly possible it doesn't work out and we never get a universal fault-tolerant quantum computer.

I think part of the reason for optimism is that there are a lot of different ways to make a quantum computer. Ion traps, linear optics, resonant superconducting circuits, topological quantum computing (if anyone ever actually discovers an anyon) and many more.

Different groups are working on wildly different directions right now, i.e. Google and IBM are doing superconducting stuff, microsoft loves topological stuff, researchers at various places are doing trapped ions and at least one company (psi-quantum) is working on linear optics.

It certainly wouldn't be surprising if several of these approaches fail to work, but the hope is that they don't all fail.