The thing I can't stand is the absolute certainty of the boosters. It's almost religious.
AI is the future. AI will do this. AI will cause that. It is inevitable. Everything is obviously changing.
They leave no room left for debate. No openness to pushback. And no evidence or proof. It just is because it is, and if you don't believe it, you're simply wrong. We saw the same sort of attitudes with blockchain and NFTs.
Any quality, useful (commercial) piece of software made with only AI code would suffice. Any successful AI-written indie game. Any successful AI-written library.
The proof should be in the pudding. Claude Code has been out for a year, and the boosters are saying they are orders of magnitude more productive, so we should have seen some kind of successful output within an order of magnitude of the time it would have taken to develop using traditional, non-LLM tools.
Where is the agent-coded Photoshop clone eating Adobe's lunch? Where is the agent-coded Quicken clone which puts the original out of business? Where is the agent-coded hit video game? All we are seeing so far are "Show HN" level projects.
The only thing LLMs seem to be speeding up is people's mouths.
AI is the future. AI will do this. AI will cause that. It is inevitable. Everything is obviously changing.
They leave no room left for debate. No openness to pushback. And no evidence or proof. It just is because it is, and if you don't believe it, you're simply wrong. We saw the same sort of attitudes with blockchain and NFTs.