why is building a large, complex and hard to produce solar panel

Solar engines have historically been simple and easy to produce relative to coal and gas engines. We've had industrial solar heating technology since the 19th century, even. Sort of the joke of fossil fuel technology, in that it's been a consequence of heavy R&D investment and industrial build out as much by choice as by any engineering advantage.

It’s improving the tech, which could eventually far and away outpaced every other energy producer.

It's been a pop science empty promise for decades. We were talking about fusion technology in the 1960s like it was just around the corner. Yes, there's an enormous amount of output in fusion technology. But it requires even more energy input. What's more - as the article downplays but is forced to concede - it is highly unstable and difficult to sustain, even at a super-sized laboratory setting.

This is a far cry from fission which can be achieved practically by accident (see the Chicago Pile-1) and is comparatively straightforward to control via mechanical methods.

Even in the event you manage to produce something approaching stable fusion at the size-scale necessary for industrial deployment, there's still no reason to believe the technology would be cheaper per watt than the indirect capture of an already running supermassive fusion system (ie, the Sun). In that case, it wouldn't outpace every other energy producer for the same reason conventional nuclear power hasn't.