We had some really interesting developments in machine learning before LLMs - they were slow but steady, every week something interesting came up. And them came

We had some really interesting developments in machine learning before LLMs - they were slow but steady, every week something interesting came up. And them came the LLMs and basically most money started flowing in this direction, for better or worse. It's hard to say whether it makes sense or not but what has been achieved already has some value for some particular tasks.