Howdy, I've done a fair bit of game research/patching/romhacking/etc in my time, and I wanted to contribute some meta-perspective to hopefully help drive peoples' spirits up a bit. Excuse the super long post, and consider showing this to your reverse engineer pals if you think it's interesting. Okay, here we go:
First off, one huge rule that's given me a lot of success over the years is that, not joking, you have to believe in yourself! It's simply too easy to sit and make up reasons that a romhacking goal isn't possible, as opposed to planning+forging your way forward until you physically hit a roadblock. You have to practice thinking positively and imagining new outcomes, even if it's unpopular or whatever. This is a major separator between folks who do and don't move the needle, in my experience.
Seriously, self-doubt will set a ceiling on what you can achieve MUCH more quickly than even your current skill level will. Skill can and will grow as you do more; you'll never regret practicing growing skills instead of so-called 'learned' cynicism!
On group sizes: I work alone on my projects usually (after coming up to speed in a tight 3ish person group as a teen) and find that it's actually better than working with a team much of the time. If you've got the brain for it, having a whole system stuffed in your head can be a lot better than the slow process of technical communication with other people. If you reach out to a group and nobody has an answer for you, you actually run a way higher risk of thinking something is fully impossible (especially because of how powerful social influence is for humans). If you run into that situation alone enough times though, you'll actually start to develop some siiick skills that let you hurdle over/smash apart entire categories of roadblocks instead. While not made by a single person, I always love to point out that Banjo-Kazooie and other high-end Rareware games of the 90s were all made on a countryside farm, without internet access!
Planning and thinking are actually /more/ important than the actual work you do with your mouse and keyboard. A lot of the process of romhacking involves wrapping your head around what you're looking at to begin with. Then you have to take time to reimagine the overall system design to include the elements you want, then finally you can actually start taking steps toward programming/implementation. Infamously, having source code often does not make this much easier in reality, as even professional programmers need time to learn how to work with unfamiliar source code. (Hex-Rays/Ghidra can often produce more modern, sane representations of scrappy 90s C code (and even handwritten assembler) anyway..!) This plan => execute process very closely mirrors the preproduction => production process that cleanly-developed games follow. Avoid overlapping core research with core development!
Remember to assess your selection of tools properly and fully-- e.g. in this case, PSX debugging and extensive logging are both trivial to do these days, even on real hardware (thru expansion ROMs like UNIROM with debugging shells-- yup, you can use your legitimate CD-ROMs!). Also make sure you take time to read old primary technical sources, and prioritize what they teach you over common/sensationalist information which might lead you far astray from the truth. Such sources can also help you understand the mindset and vibe of the programmers, which is useful for understanding peoples' design choices.
Polyphony did a legendary job on many, many games. And, they are humans. All programs are designed and laid out by humans, in a way that makes sense to those humans. Even CPU instructions have a human element to them, being that the CPUs themselves are also structures designed and laid out by humans. Snarky points about future AI-designed systems+chips aside, when it comes to RE, it's important to remember each layer of complexity was designed by people like yourself-- not some sort of inscrutable heavenly being!
If you break things into pieces over and over you'll end up with a stepladder directly to your goals, when it comes to these things. The only time you should pause for cynicism is if you legitimately can't find the path between two of the steps toward your goal, or if you need to second-guess plans that didn't work as expected.
I'm personally in the process of circling GT1 and kicking the tires so to speak, sizing up if I want to try going after some minor goals like replacing the soundtrack (has already been done but sounds like a fun drop-in point nonetheless), messing with compressed code, etc. To me, making a version of a game that's slightly more to my taste is a way of having a no-frills direct talk about technology with its creators, a talk based in absolute ground truths about the works that they've created. A type of interaction and introspection on sources more accurate than even the programmers' now-27-year-old memories.
Speaking technically: There's a lot of room to jam code into almost any PS1 game, you just have to think out of the box and stay in sync with how your particular title works. As an example, there's like several entire kilobytes free before the usual PS1 executable load address. Also GT1 is overlay-heavy already, so your flecks of patch code can be dynamically-loaded overlays too, no? For re-injection of modified stuff, you can often just tack data onto extra LBAs at the end of the disk and then patch up relevant code to look at that new location. (if otherwise-valid data became larger after recompression, for example). There's no need for a full decompilation to achieve most goals, just careful patching and strategy to live alongside the existing code and data. (Not that a full decomp wouldn't be cool!) I don't want to go on too long about this since I generally wanted to talk shop and theory (and am still learning GT1's structure), but figured it'd be good to offer some loose tech theory.
You don't have to perfectly obey the existing programming to achieve your goals successfully, you just need to work politely and subtly alongside it. Final Fantasy VIII was practically hacked into English using GameSharks since Square JP wouldn't trust others with their code. Game porters in the 80s were often given nothing more than a floppy/cart with the existing compiled game on it to work with-- and often achieved direct ports nonetheless. Someone has always made ends meet in a harder situation, or at a grander scale. Sure, OP's examples simply aren't all achievable, but their mind is actually 100% in the right place.
At the end of the day, the whole game is a ~5.5 gigabit volatile hunk of data. It's not made of hardware gates and wires, nor gears pistons and belts. There's a finite number of bytes with a finite number of purposes each, being executed on a very average computer that tens of thousands of people all separately learned to program. Each piece of it makes human sense, and systems can be designed alongside each piece to reengineer them as needed.
Phew, that's all. I just get bummed out when people say broad swaths of things I've literally achieved solo are, erm, not possible? Like, you can do it! I believe in you! And fans of the games deserve the positivity too! OP is very on-point keeping his mind open, and the combination of that with technical ability and confidence is one of the most powerful forces in the universe. Heck, it's the force that created the games themselves! ^_^
