In 2022 I was toying around with OpenAI's RL Gym, right when the first non-instruct GPT3 model came out. I was thinking about getting into ML a lot more, but hesitated. Before that it was 3D printers, mechanical keyboards, drones, etc. All of these have exploded, and while they are still very interesting, I do love my Browns and manage Prusas for my local hackerspace, they have just, for the lack of a better term, industrialized. I'm also now in a position where I have time and money for it, not like when I was 15 and rating Ender motherboard upgrades I knew I'd never buy.
Right now I'm making a chess engine, but that's already a solved problem. There's also biohacking, and while designing chips to go into my body is really interesting, I only have one, and don't want to push it too far. One promising idea is a kind of 'Personal Computer 2', where people try to innovate HCI, and while I really like that and do have some research ideas, I'd like to explore a bit more before delving deep into it.