The "it's so over / we are so back" cycle in AI has happened about 40 times in the last two years. Every two weeks, there's a new model and an updated benchmark, followed by the latest discourse. If you'd kept building with something like Opus 4.5 this whole time and reassessed every six months, you'd have been fine. I'd be really jealous of the people who've actually managed to do that. The rate of change is psychologically disruptive regardless of whether you think AI is good or bad. You…