A quick question, what accent do you have? As an Aussie I have real trouble speaking naturally with most speech to text software, open or closed source. I feel like the guys in that Scottish sketch show in the voice activated elevator. I sometimes use voice dictation for my notes for work and I spend almost as much time correcting as I do speaking.
That said, I have found a perfect solution. I can get well over the 95% correct mark by simply using an English or American accent. I can do both fairly well and the speech to text has no complaints. I imagine someone from Boston would have a tonne of trouble being understood, as would a Welsh person, but pretending to be a Californian or similar can help immensely.
I would love to find something that can be trained by my speech like Dragon Naturally Speaking used to be. I used that in the early 2000s and at first it was awful, but training it for a few hours really did offer a noticeable improvement, and ongoing use continued to improve further. My computer died and I lost all the trained data, so I never went back, but if I could I would definitely do that again.
I think this should be a thing. At the same time, I would also want something similar for funding or platforming fascists, but that is unlikely to end up being done. I think a simple tag, the [AI] one would work, is the best current solution. I think extra detail in the post is a good thing to do, for example AI assisted documentation, AI assisted bug finding, AI assisted vibe coding. They are all different and have different effects on the product and community. If someone uses AI to find bugs in their own code I am all for it, that is a great use if it. If they use AI to write their login system I am not keen at all given the likelihood of intense security issues and the low likelihood that they will ever fix it.