The big companies fighting to win gen AI have forgotten how to build good products. There seems to be enormous confusion about AI applications — few seem to be willing to commit to a singular focus or application, instead trying to be everything to everyone. The tried and true way to build something big is to bootstrap a killer app that solves a real problem and then expand horizontally to other problem spaces. These companies are trying to do too much too soon (probably because they have too few financial constraints). A lack of focus.