I actually agree with them. (Both are very smart guys as well.) I use AI heavily too. I think using it well is a skill unto itself though and that is still very much getting figured out.

Two things can be true. (And usually are in hype cycles like this.)

People probably think I sound like schizophrenic on the topic. On one hand I am finding massive utility. On the other, I am telling hype drunk vibers to sober up.

I think the AI resistance and AI mania sides are both profoundly wrong. AI is new tooling and my guess, based on observation personally using it and experience with multiple tech cycles, is that a new, more powerful elite engineer will emerge.

Management, low end specializations and people who won't adapt will take the hit. Paper pushers, bit shovelers and people who just want to tap out. A lot of koolaide drinkers too.

That's what always happens. As Cory Doctorow recently put it, the hype is always shaped by how much the "bosses" hate engineers and salesmen selling to that. I find his progressive priors distasteful, but he is not wrong. It happens this ways every damn time.

The reality check stands though. Where's the beef? It's a sincerely curious question. Can anyone point to a real piece of software? If you can clone Bloomberg in hours, where's the product? If not, there is a huge productivity illusion that is part of how we currently perceive AI in practice too, and we need to understand what that is.