I appreciate the concreteness and perspective you bring to the subject. Hinton’s hyperbole can be efficiently challenged by attacking the myth that AI displaces jobs. As you make clear, AI displaces discrete tasks.

Beyond this surface level endorsement, I look forward to a deeper analysis of some of the forward-looking trends in AI. For example, you discuss barriers of diagnostic reasoning and trust/interpretability. These problems dominate research programs and are fast-advancing. As highlighted in another comment, when your time-horizon is decades (a person’s career), criticisms in this area may be relatively short-lived.

Most seriously, I fear your piece ends on another common myth: AI augmentation. Can’t we all just get along? It’s an argument that is in direct opposition to the opening frame, that AI will relentlessly erode these tasks. The tasks remaining for people will be reduced to a relatively smaller number of elite practitioners, who are made enormously more productive. Drawing an economic argument of increased demand only begs the question of market satiation. Just ask the journalists…