I think the concepts are inevitable, not so much the specific implementations. PCs were an inevitable stop on the fairly standard adoption path from many workers to one machine (Mainframes), 1:1 (PCs), many machines to one worker (I have at least 10 computers within arms reach right now). IBM/Windows/Apple weren't inevitable, they were just manifestations of that. ICs weren't inevitable, but commoditized computer parts were. TCP wasn't inevitable, but a lingua franca for networks was. LLMs weren't inevitable, but AI is.
To your overall point though, and to the contrary of the type of thinking you're critiquing, the timing of these things is not inevitable. Computers didn't have to happen in the 50s, they could easily have waited 50 or 100 years if we didn't have things like wars or other technological breakthroughs that enabled them. For AI we might be stuck on incremental improvements on LLMs for a generation, or they might be obsolete in 5 years. They will be replaced by something better at some point, but confusing (intentionally or not) inevitable with soon is where the hype proves itself hollow.
To your overall point though, and to the contrary of the type of thinking you're critiquing, the timing of these things is not inevitable. Computers didn't have to happen in the 50s, they could easily have waited 50 or 100 years if we didn't have things like wars or other technological breakthroughs that enabled them. For AI we might be stuck on incremental improvements on LLMs for a generation, or they might be obsolete in 5 years. They will be replaced by something better at some point, but confusing (intentionally or not) inevitable with soon is where the hype proves itself hollow.