Looking at it from an objective/historical perspective, when he first got on his high horse agents weren’t a thing, MCP was in its infancy at best, and half the code generated didn’t compile from frontier models.

From there everything split into two factions, which I’ll dub as believers and non-believers. From there, it has entirely been a cult following despite the evidence showing that models and agents have legs for software development.

FWIW my post history would show my extreme skepticism, and to an extent I still am. I think the real power of models isn’t the model at all but the harnesses.

Either way I think he lost the plot and runs on vibes himself. I also think this whole AI movement is going to have their 2000/2008 moment before the phoenix rises from the ashes.

I don't agree with the believers vs non-believers dichotomy. That makes this look like a religious war and it shouldn't have to be like that. To me AI is a very useful tool, no more no less, it isn't a 'make a wish' machine and it isn't a silver bullet for all of the issues that have plagued software so far. But when properly applied it's quite useful. "Non-believers" would have to be people that have yet to actually use AI, just like you can probably be a non-believer in peanuts until you've seen them, and most people that I know acknowledge the reality of AI tools and their use cases.

All tools, AI included, have a business end and as long as you point that away from yourself you will be fine. But to properly apply it (rather than as a faster way to make a huge mess) takes discipline and being methodical. It was never different.

what's it with people who think a bubble is literally destined to happen because of some religious belief that history repeats itself? I find it fascinating to understand the theory of mind of such people. So much confidence.