The bet against AI was quite safe, until it wasn't
There was a firm, implicit conviction in business that things that AI would decimate still had a future because AI was still a long way off.
Like the time traveler who went back in time but never bet on things that actually happened, those that had the most faith in the developments we see today before they took off never shorted the things AI would undermine, at least as far as we know, but maybe the reason we don't know is because AI hasn't properly undermined them yet? The reason is simply that ai made with the current techniques (ie, llms) will never understand anything, and cannot replace humans. The way they work is simply the issue here. The hype was (and still is) grifter techbros hunting the big money.