> The core here is, that if engineering is going, then law is going, marketing is going and a lot of other professions are also going
The difference is reliance costs. We go to a doctor or a lawyer instead of a book because the risk of mistake is too high.
But development work can be factored; the guarantees are in the testing and deployment. At the margins maybe you trust a tech lead's time estimates or judgment calls, but you don't really rely on them as much as prove them out. The reliance costs became digestable when agile reduced scope to tiny steps, and they were digested by testing, cloud scaling, outsourcing, etc. What's left is a fig leaf of design, but ideas are basically free.
This seems like a kind of arbriatry dichotomy to setup to support some arbratary argument.
Let me put it like this: when LLMs write system scale software that is formally verifiable or completely testable, I promise you that you don't trust your doctor more than you trust the ai.
Already now we see indications that diagnosing is better without human (doctors) intervention.
Fast forward ten years with that development and nobody dares to go to the doctor.
The difference is reliance costs. We go to a doctor or a lawyer instead of a book because the risk of mistake is too high.
But development work can be factored; the guarantees are in the testing and deployment. At the margins maybe you trust a tech lead's time estimates or judgment calls, but you don't really rely on them as much as prove them out. The reliance costs became digestable when agile reduced scope to tiny steps, and they were digested by testing, cloud scaling, outsourcing, etc. What's left is a fig leaf of design, but ideas are basically free.