I don't really know of any ai researchers in our extended network out of some dozens who've managed to be taken very seriously without being colocated with other top researchers, so without knowing more, it still seems moderately likely to me that the best plan involves doing something like earning while practising math, or studying a PhD, with the intent to move in 2-3 years, depending on when you can't move until.

Otherwise, it seems like you're doing the right things, but until you put out some papers or something, I think I'd sooner direct funding to projects among the FLI grantees. I'd note that most of the credible LW/EA researchers are doing PhDs and postdocs or taking on AI safety research roles in industry, and recieve funds through those avenues and it seems to me like those would also be the next steps for you in your career.

If you had a very new idea that you had an extraordinary comparative advantage at exploring, then it's not concievable that you could be among the most eligible GCR-reduction researchers for funding but you'd have to say a lot more.