why it’s too early to claim what job sectors will be overwhelmingly different in the future

At least based on the path we’re on right now, AI will be best harnessed by people who understand the domain they’re working in (or problem they’re trying to solve). And thus, AI will ultimately augment how we work vs. replace it.

With even a small and decreasing hallucination or error rate, you are ultimately bottlenecked by your ability to give AI agents the right context to work with, review the output of AI, and incorporate its work into a broader set of tasks. The people who understand their field the best, and have high judgment, are in the best position to do this over non-experts.

A non-expert can learn about new solutions to problems, prototype ideas faster, get quick answers to questions, or become an expert far faster if motivated. But the experts will be able to actually produce full work faster because they will know how to veer an agent on the right track, how to review what work is wrong and how to fix it, and so on.

This is also why it’s too early to claim what job sectors will be overwhelming different in the future. It’s still much better to go very deep in a field and become an expert at a category of work. The best lawyers, engineers, scientists, marketers, or designers will be the ones that get the most leverage from these tools in the future.