Well, reworking Bing and Google for a ChatGPT interface is going to be massive hardware and software enterprise. And there are a lot of questions involved to say the least.
Where will the software engineer come from? We're in a belt-tightening part of the business cycle and FANGs have a pressure not to hire, so you assume the existing engineers. But these engineers are now working on real things so those real things may suffer. Which brings actual profits? The future AI thing or the present? The future AI is unavoidable given the possibilities are visible and the competition is on but a "shit shows" of various sorts seem very possible.
Where will the hardware and the processing power come from? There are estimates of server power consumption quintupling [1] but these are arbitrary - even if it just doubles, just "plugging the cords" in takes time. And where would the new TPUs/GPUs come from? TSMC has a capacity determined by investments already made and much of that capacity is allotted already - more capacity anywhere would involve massive capital allocation and what level of increased profits will pay for this?
[1] https://www.wired.com/story/the-generative-ai-search-race-ha...