I don’t think this is a balanced view on LLMs. I particularly found it weird to liken LLM creation as akin to the global race for a covid vaccine.

There are a lot of assumptions about the pace of improvement (which is already tapering rapidly). There is no consideration for the inherent biases that are built into these LLMs. LLMs make shit up all the time. They don’t think. They can be easily tricked and manipulated at multiple stages in the training and use.

I think LLMs will change things for a lot of industries, absolutely, but there is a limit.