100%. We also need to accept that:
writing strict implementation docs > writing SDKs, in the agent era.
Why? Because abstractions built to hand-hold humans are becoming bottlenecks for AI. Agents don't need your boilerplate code; they need raw context.
end-to-end testing > unit tests, in the vibecoding era.
A massive, almost entirely agent-coded refactor passed all unit and pre-merge tests but broke a critical feature.
It was only caught due to my own excessive paranoia making me run end-to-end tests before the prod deploy.