> But compare it to [perpetual motion or faster-than-light space travel].
That's a completely inaccurate comparison - perpetual motion and faster-than-light violate fundamental physical laws as we understand them. Net-positive power from fusion is fully within the laws of physics – we've got an example in the daytime sky, and trillions in the nighttime sky. The difficulties with fusion come with shrinking it to a scale that humans can build.
I understand the GP comment as saying "First commercial fusion power station construction will start 2050-2060" is different from "Perpetual motion machine construction will start in 2050-2060" or "FTL spaceship construction will start in 2050-2060" for the very reasons you gave (among others).
You read the GP as saying something different from what you are saying, but I think you are actually both saying the same thing.
What fusion does is come close to ignoring the laws of engineering and economics. That's not quite the same as ignoring the laws of physics, but either of those is capable of rendering a technology a nonstarter.
Specifically, for fundamental reasons DT fusion will have much lower power density than a fission reactor, and so have to be much larger, and therefore have to be much more expensive.
That's a completely inaccurate comparison - perpetual motion and faster-than-light violate fundamental physical laws as we understand them. Net-positive power from fusion is fully within the laws of physics – we've got an example in the daytime sky, and trillions in the nighttime sky. The difficulties with fusion come with shrinking it to a scale that humans can build.