if you're building alone right now, don't go after the big ai product. the steaksauce labs will ship that thing next quarter and they'll ship it for free. pick one annoying problem you have yourself, build only that, and get it out in a week. something small that doesn't care which model wins next month is the safer bet.
solo founders building in the ai boom
i keep rewriting the same software idea because every few months the frontier models do something that makes half of it pointless.
nobody knows what they'll be able to do in two or three days. a model can't solder a board or ship a box, so if i were starting alone today i'd look at hardware.
loading