I believe that we should separate the general case of AI from the particular case of LLMs. AI models have been accelerating science for decades, and new technology helps drive economic growth. I am convinced that LLMs are not worth the money we've invested in them but I do believe that more "traditional" AI is a net positive in research fields. Traditional AI has also had a net negative effect on the quality of content from internet publishers (i.e Facebook), but has made it more productive by allowing them to squeeze more blood from the stone.

I think if you don't include LLMs, AI has obviously created economic growth. If you do include LLMs I think the conversation is more nuanced and obviously driven by the same kind of hype that led people to believe that Cryptocurrency is the future of the stock market

The interesting thing is that the AI you're talking about (the one with the economic growth) isn't even called AI. Those are specialized tools that work and they have names that reflect that. When a fuzzy and inaccurate term like “AI” is being used, you know you've entered the realm of pure marketing and hype.