I think automation still represents a threat to fields like CPAs, just not an existential one. Historically the specialized/knowledge fields like accounting/medicine/etc. have been able to provide jobs even to those who graduated at the bottom of their class and/or were just marginally proficient just because a lot of the work is pretty simple - the equivalent of $30k tax returns or diagnosing eczema in my example. Extremely routine tasks like that in accounting/medicine can likely be automated, so if you are a barely competent CPA or doctor then you are very correct to be worried about automation. But it also means that because AI cannot realistically be programmed to handle the extreme edge cases like ones I mentioned in my post you quoted, if you're a really, really good doctor that your job is basically immune to AI. That means that you need to be even that much better, more competent, more creative to still have a job in those fields once all the low hanging fruit tasks have been automated away.
That's going to mean a few things - to an extent, job satisfaction will go up because a lot of the menial, mundane, and repetitive tasks will be taken off your plate. You won't have gone to medical school plus years of residency and such just so you can spend 1/3 of your time treating people with nothing more severe than allergies or muscle sprains, the "AI doc" or "last in medical school MD" or lesser trained workers can handle that stuff. It means less time needing to issue and monitor relatively simple and routine orders to nurses and techs for post-surgery care and such, the AI can handle most of that. It means the tasks that are going to come to you are going to be much more complicated, require much more investigation and problem solving, and (perhaps) be more fulfilling yet also demanding. There's going to be little to zero "fluff time" in your schedule and every moment is going to require you to be at peak performance. The future doctor is going to need to be much more like House from the TV show (where the plot describes contrived but hugely complicated medical issues that require a lot of skill and creativity to solve) and a lot less Marcus Welby doing a simple blood test to find you have anemia.