As a low income indie dev without the academic credentials to get past HR gatekeepers for stable full time work, I can't help but feel studying ML is about as wishful as creating a hit game. I have started studying ML this summer, and although I find the NLP applications really interesting since I have a social science background that exposed me to some literary theory and linguistic ideas that overlap with NLP at times, I think making a living wage doing it is just as much of an unrealistic dream as writing some killer app. The jobs all require PhDs, and the data science competitions have literary thousands of people with PhDs and industry experience competing for five figure prizes. When one is poor with no prospects these kind of pipe dreams feel so good to get caught up in as that sweet haze of hope numbs the critical thinking, but in a clear moment it looks like the ML gold rush is exactly the same as every other tech hype. That's not going to stop me from geeking out on PyTorch and trying to wring sentiment out of blocks of text or whatever, but I won't be able to honestly write a blog post about it in a few years asking if I wasted my time. I already know the answer.