Some jobs require more than just intelligence--for example, nurses, chefs, judges, reporters, teachers, caregivers, therapists, social workers, hair stylists, coaches, paramedics, firefighters, police officers, electricians, plumbers, mechanics, clergy, trial lawyers and diplomats.
What is the common thread?
These jobs depend on some combination of physical skill, human connection, trust, judgment and accountability, and such jobs are out of reach of AI.
Doctors, nurses, therapists, coaches and reporters will definitely see impact. AI is already being used for all of these things, although probably not very effectively/safely. In the near future, better tools will come out that fix the worst drawbacks. All of these will still have use-cases for a human, for example in very difficult cases, or of course when a human is needed for the actual treatment, but at the margins, AI will chip away at them.
If predictions about robotics are accurate, things will get even more difficult in the next ten years as humanoid robots mature.
Automation and robotics are also surging rather silently in the shadows as AI is the lead that’s getting attention. And AI happens to be congruent with those 2.
The other common thread for most of what you've listed is, unfortunately, comparably lower pay.
Arguably, software has times that require physical presence, human connection, trust, judgment, and accountability.
It won't be tomorrow, but I'm not so sure blue collar technicians (plumbers, mechanics, electricians) are far behind: once parts become modular enough, with sufficient sensor data, a cloud system at the manufacturer for your car, boiler, or home lighting system will be able to give you step by step instructions for how to replace and repair items, and will ship you the replacements and tools needed.
Of course the elderly and frail and people too afraid to touch something lest they break it will still call an expert, and original build and installations will need those trained, but I could see vast swaths of less technically skilled people opting for a go at routine repairs which could really hollow out the demand.
I think ultimately this is a flawed understanding of the incentives at play.
None of these end up really being out of reach of AI if Anthropic can justify this valuation. They're just slower to roll out.
Is software engineering really that different? The most effective devs are those who don't just have a vision but can work across the org to generate buy-in and enthusiasm to make that vision a reality. This very human process, driven by relationships etc.
AI is already replacing some reporters, therapists, and lawyers. One might argue that it shouldn't (an AI therapist is a bit of a scary concept) but it does. And they're only getting better.
Oooh it won't be black and white: It will start with 1 helper robot per nurse, yo help her in the repetitive or heavy tasks.
Then they will be able to do more and more, and 1 nurse will be able to do the job of 4-5 2020 nurses.
Until that nurse is a dude from Mexico [where im from/based] giving instructions to all the robots.