There seem overwhelmingly many more ways to earn a useful surplus of money than there are ways to usefully research IA, so any given person (i.e. you) is more likely to be able to learn a skill from the large set of {skills which could earn surplus money} than to learn a skill from the much smaller set {skills useful for advancing the state of IA research}.
Particularly since your claim of "no skills already developed" ignores your ability and willingness to write, and your having a blog and associated skills, both of which are things some people earn plenty from.
There seem overwhelmingly many more ways to earn a useful surplus of money than there are ways to usefully research IA, so any given person (i.e. you) is more likely to be able to learn a skill from the large set of {skills which could earn surplus money} than to learn a skill from the much smaller set {skills useful for advancing the state of IA research}.
I already have a path laid out for IA research. Nobody's actually doing focused, systematized IA work; it's a really high value target right now, and it's something I'm already specialized for. In gen...
One method for increasing high utility productivity I thought up was choosing a specific well-defined answer for the second half ("Why am I doing it?") and consistently checking to see if the answer to the first half satisfyingly aligns with the second half. For example, if I'd checked myself an hour ago, it'd be "I'm learning to program because I want to maximize the probability of FAI development." Ideally the second half would be related to a 'something to protect' or 'definite major purpose' that stays constant over time and that you want to be consistently moving towards. If you're already good at noticing rationalization this technique might work to induce cognitive dissonance when engaging in suboptimal courses of action. (Whether or not inducing cognitive dissonance in order to make yourself more productive is likely to work is open to debate. I suspect P.J. Eby would thoroughly disagree.) I'm going to try this over the next few days and see if the results are any better than how I've been doing recently. I am at a relative productivity high point right now though, so the data might not be too meaningful. I encourage others to see if this method works.