how completely mired in stupidity the forum software is, and they're too incompetent to fix it because the guy who cobbled it together in 1997 left and now trying to unfuck their site without a complete rewrite of everything is not feasible.
To be fair, a lot of places are like this for some time after starting up.
They have very limited money when they start so they'll either hire a very small team of cheap in-house developers (I try not to stereotype, but I've observed a pattern where a lot of those developers implement the whole thing in php) to implement their website or they'll outsource the coding to a small company, specialised in making websites on the quick. Either way, the rule of the game is to implement everything yesterday on a shoestring budget.
Even if the developers involved were top notch developers (they usually aren't... most of them haven't updated their skills enough to use anything better than php afterall and often will lack awareness of sound design patterns/standards like using proper REST for http apis even over a decade after such patterns have become widely known), the time constraints pretty much guarantee that the resulting system will suffer from significant technical debt right out the door.
From there, they'll hopefully start making money and the amount of income they generate will dictate the future quality of the system. If they make tons of cash, they'll be able to hire one or more teams of skilled developers with more time to improve the code (at that point, they'll either do a complete rewrite from scratch or slowly refactor away the legacy codebase, depending on how much of a mess things really were and how much resources they can invest).
If they don't make as much money as originally anticipated (keeping in mind that their primary goal is to turn up a healthy profit, they won't hire a crack team of developers if it will eat away a significant portion of their profits, even if they could do so while remaining green), then technical debt won't be repaid (or will be repaid extremely slowly) and the system will probably remain cr*ppy for the remainder of the company's lifespan (after which it will, thankfully, be shelved).
Given that we don't have access to GOG's financial information (and frankly, I haven't taken a look at their codebase, though the fact that they are still working in php is definitely not a good sign), its hard to know exactly what is happening there.
Thee's a slight flaw with what you said. It's a bit of nuance, but it's really worth mentioning because it leds to how people fail to see the difference between a publicc and private company. A private company can afford (and will so if they believe the project is worth it) to hire craack developers. A public forporatin, however, cannot, due to legal constraints.