I can't believe these clowns put "veterinary assistant" on the list of jobs where at least 50% of the tasks could be feasibly done with an AI agent. Do they have pets? It's a highly physical job - they have to take care of animals and administer medicine. Administrative stuff is maybe 20% of it.
"We trained our robot on 50,000 dogs and it has superhuman pet-weighing abilities! Sadly we forgot to train the algorithm on rabbits, and in a real-world deployment Floofkins was killed instantly."
Likewise with property manager, event planner, food service manager: these jobs involve a lot of squishy real-world common sense which is notoriously hard for AI even with big data, and the data hardly exists for this stuff. I am sure LLMs can help with the paperwork. But the valuable skills for these jobs are essentially orthogonal to the recent progress made in AI.
The only explanation I can see for this idiotic list: they took an enumeration of the tasks for these jobs and measured 50% by-item rather than by 50% by-time. I could see a vet assistant having 50% of tasks being like "schedule appointments" or "file paperwork." But instead of saying "wait these conclusions are absurd, how is an LLM going to feed a cat?" they used their tech/finance-bro's sneering contempt for lesser jobs to rationalize a dumb conclusion.
