Laser inertial confinement is a commercial dead end. It's only continued as an inertial funding exercise, reusing nuclear weapon simulation facilities long after supercomputers took over the modelling.
[0] https://en.wikipedia.org/wiki/National_Ignition_Facility#/me...
Furthermore, supercomputers do not replace these facilities, only complement them, as modeling a system solely in computers is a good way to find you have drifted away from reality.
Sure, but that's also why they don't care about the inefficiency of the lasers themselves, which we already know how to improve upon significantly.
Which means…
> Still two orders of magnitude to go, based on 300MJ inputs. Not gonna happen.
Might happen.
Not guaranteed — even good lasers aren't great, though apparently 60.5% wall-plug efficiencies have been demonstrated[0], and even if the laser part was perfect there would be a lot more steps from this to making a power station (or a fusion rocket) — but it's certainly in the realm of being worth considering.
[0] still a loss overall, assuming 150% gain from laser in to fusion out: https://ui.adsabs.harvard.edu/abs/2020JSemi..41c2901R/abstra...