surprised at how many people on record with strong "AI isn't going to replace job X" views continue to focus exclusively on the current tech rather than on its first and second derivatives.

the gap between GPT-3 and GPT-4 was < 2 years. The _first_ LLMs (BERT and GPT-1) are only 5 years old. it takes some _very_ pessimistic (or, depending on your views, optimistic) extrapolation to draw conclusions about what AI will be able to do even a year from now based only on current capabilities