"In 1988, the roboticist Hans Moravec noticed something strange. Computers could already beat humans at chess and prove mathematical theorems, but they couldn't reliably pick up a coffee cup or walk across a cluttered room. What's hard for humans is easy for computers, and what's easy for humans is hard for computers. That observation, now called Moravec's paradox, has become the single most important fact about which jobs AI takes over and which it doesn't."
"Document review, contract drafting, basic medical diagnosis, financial analysis, code completion - the cognitive work that requires years of education and pays six figures - turns out to be relatively tractable for AI. Folding laundry in a stranger's house, repairing a leaking pipe behind a finished wall, calming a frightened toddler, diagnosing a strange noise in an aging Camry - work that pays $25 an hour if you're lucky - turns out to be extraordinarily hard."
"Most coverage assumes the cheapest jobs automate first. The wage hierarchy of automation is inverting. The trades face a different transition than knowledge work: Augmentation, not replacement. A master electrician at 28 will likely have a more stable career than a marketing grad of the same age."
Moravec’s paradox describes how computers handle tasks that are hard for humans and struggle with tasks that are easy for humans. AI can perform cognitive work such as document review, contract drafting, basic medical diagnosis, financial analysis, and code completion, even when it requires years of education and high pay. In contrast, physical tasks like folding laundry, repairing plumbing behind finished walls, calming toddlers, and diagnosing mechanical problems are difficult for AI and often pay far less. Automation therefore does not follow the assumption that the cheapest jobs disappear first. Trades face a different transition than knowledge work, moving toward augmentation rather than replacement. A young master electrician is likely to have a more stable career than a marketing graduate.
Read at Psychology Today
Unable to calculate read time
Collection
[
|
...
]