There’s two issues this leads to:
1. The use cases you value it for really don’t have enough ROI behind them for the cost of building and running the models, even with economies of scale. I often make the analogy of highly advanced robots being very useful but economically inaccessible to most. AI is effectively subsidized in an unsustainable way right now.
2. Gen AI has to be hyped as a panacea to try to mitigate the first point, so it’s shoehorned it into many ill-suited use cases.