The barrier to entry to generating code may be "I can think", but the barrier to entry for solving hard, distributed/multi-faceted engineering problems still remains quite high - agents can't really do this still to a decent level of efficacy reliably.

The progress models have made in the last 5 years aren't convincing me they'll bridge that gap too soon, although I can see how some people are convinced by how decent agentic harnesses make things. I know it's really easy to get very hyped with the current state of the technology, but try to have a bit of skepticism.