AI is no good at thinking yet.

It isn't very good at new solutions, yet.

It isn't good at being novel and original, yet.

But those features are approaching rapidly.

A combo of introducing some randomness into its output, combined with coded feedback paths for it to examine its own output, and some other structural tweaks we haven't discovered yet will give birth to a general intelligence.

And this general intelligence structure is found, it will get copied by tons of folks almost instantly (if the corporation that figures it out shares it).

The world will start to be very different once that happens. It is starting to be very different with the agentic frameworks.

So in this one case I think you're wrong and I think AGI will make a big difference very quickly, even considering the sheer mass of society and its inherent inertia against change.