> I wonder why everyone skirts around the obvious here: the NIF's task is weapons research, and for that, fusion reactions which last nanoseconds might be enough, but there is no way this method can be scaled up so it's actually a cost-effective way of producing electricity...
I think you answer your own question.
NIF does a lot more than weapons. Ignition is also a different problem than sustaining a fusion reaction, but both are important aspects of the fusion energy generation problem. You can think of this like the difference between static friction and kinetic friction. You would never try to generate power by performing tons and tons of ignitions. You generate power by doing the ignition and then sustaining the reaction. There's even people working on setups just like this. You can see how ITER works here[0]. You can modify this procedure to have laser based ignition (or some other form) and then feed this into the toroidal reactor. There's a bunch of ideas out there but I think this is the most straightforward.
As to a more systematic approach to answering "why" and why this type of language is pervasive regardless of the domain, is because a lot of people have a difficult time differentiating TRLs[1]. Or maybe even more bluntly, research from products. I think it won't be hard to find high rated HN comments on research works that come out of universities who's critique is that the process is expensive and doesn't beat current industrial processes, but this is a grave misunderstanding of how technological progress happens. It's overly dismissive. Technology often progresses by iterative discontinuous S-curves, which at a zoomed out level looks more smooth. So new technology often starts off way worse but the key part of this is the theoretical maximum. A good example of this might be in batteries or solar panels. People are highly dismissive of any technology that is yet to become a product (TRL 6-7-ish). Which it is good to not jump to assertions that a low TRL result means new products, but it's the same type of error, just in the other direction. Either way, critical context is being ignored.
I think if we more appropriately contextualized technological development then we wouldn't have these issues. I don't like them either tbh. I'd rather be honest. But truth has bounded simplicity and lies don't. Since a lot of people ignore nuance it creates an incentive system to exaggerate or leave out important details. Frustrating, but understandable. Hopefully understanding can help us to disrupt this systematic incentive system and instead promote more nuanced and honest discussions. But I think we need to recognize expertise does not transfer between domains well and discourage claims and comments that lack nuance. But nuanced and accurate comments require more words, so we might just be fucked and this is just a pipedream. Some people think we can't rely on humans being nuanced and intelligent creatures but I'm not convinced and more convinced these are just products of environmental pressures. I guess we'll see.
[0] https://www.youtube.com/watch?v=5tH4obUsY64
[1] https://www.nasa.gov/directorates/somd/space-communications-...