There has been some discussion about whether artificial intelligence replaces entry-level work or just does the same work ten times faster. Both views miss a very basic cognitive notion: Intuition develops as a result of friction.
Entry-level repetitive work used to do more than execute low-value actions; it provided a form of cognitive apprenticeship which built up the necessary mental models. The experience of spending many days tracking down the little memory leaks or cleaning up messy data environments was what formed the experience of learning what failure modes are and what signs of suboptimal architecture look like.
The moment when entry-level talent skips that necessary friction and settles into jobs which involve just editing prompts and checking attitudes, vital skills degrade. There are more blind spots because there is no experience "scar tissue" which would allow for catching the false but believable information before it goes to production. The context becomes superficial: people check one isolated function, but don't think about the emergent behavior of the whole system under pressure. Debugging muscles become weaker because of the reliance on the model for the solution of the problem which does not happen because of the lack of the basics.
So, what will be the expected results in the next five to ten years as current senior engineers, system architects, and domain experts retire?
[link] [comments]