Eh, I think that's just the state of contemporary "influencer"-heavy discourse. The similarity in how people talk about the two fields has more to do with the people doing the talking (they're the same people) than it does with similarities between the fields themselves.

While LLMs are still new, ML itself is by no means a new or unproven technology. Recommendation engines built on ML are, for better or worse, the backbone of the average person's online experience. Fraud and spam detection, built on ML, are baked into most of our digital utilities (email, payments, etc.) How many apps do you use that involve ETA predictions? Because those are all built on ML. Image recognition, speech-to-text, even the camera on most smartphones--it all relies on ML.

Even looking purely at LLMs, they've reached a point of adoption that blockchain never did. ChatGPT alone reports over 200 million users per week. Google search, possibly the most ubiquitous product in the history of the internet, uses LLMs as a standard part of search.

You may have criticisms of the technology, you may think it is overhyped or that companies are making a mistake by leaning into it, but it's difficult to rationally dismiss it as a purely speculative hype bubble when its usage is so widespread.