Speedrunning is the type of thing that was kind of interesting back in 2014 or so, now it's just stale. I stopped watching Mario 64 runs because it's like, people competing to shave off 1 minute TOPS, it's not really all that interesting to watch, it's more interesting to watch a video of someone that understands code explaining how a certain exploit works, why X is in the game etc.
Yeah, the technical aspects of it are still interesting, and I can even still find entertainment in seeing somebody finally succeed in cutting 8 frames off a run after trying hundreds or thousands of times, because there's a wholesome charm to the kind of happiness they feel at that moment. And there's still something fun(ny) about watching lots of autists come together to try to break games with controllers. It's just that the in-person stuff stopped being fun when it got flooded with trannies and the requisite "no fun allowed" rules immediately followed.
But the code and techy side of it is much more interesting to me, largely because you end up learning so much random shit about games (and their original hardware) that's only being explained to help understand a specific glitch or whatever, but you wind up finding out wild shit like "the neutral zone in Yar's Revenge isn't actually random, but rather a visualization of every other 4-bit word of raw data read from a part of the cartridge ROM (because it was faster than generating random numbers), and the programmer who did it had to fight tooth-and-nail to convince his manager that he wasn't 'giving away the game code' by displaying it that way."
For programmers like me, that kind of out-of-the-box thinking is inspiring and gets the brain grinding in very satisfying ways. That particular kind of innovation isn't necessarily
useful anymore, but it serves to remind me that there is almost
always another (potentially better) way to do something, no matter how tight and lean you think your algorithm is. Somehow that specific example also helps keep my head screwed on straight in terms of "premature optimization" for some reason, too. Dude writing Yar's Revenge didn't bother fucking around with it until it became necessary (running out of cartridge space and seeing sub-par performance), and by the time it did, the solution practically jumped out at him because it made a lot of sense.
