Ten thoughts on AI that are economically useful

There are multiple great truths in this post. I'll quote a couple of them.

"...AI moves all costs to prompting and verifying. Basically, today’s AI only does tasks middle-to-middle, not end-to-end. So all the business expenditure migrates towards the edges of prompting and verifying, even as AI speeds up the middle. [...]. Today’s AI is not truly agentic because it’s not truly independent of you. The current crop of agents can’t set complex goals, or properly verify outputs. You have to spend a lot of effort on prompting, verifying, and system integrating. That just means the smarter you are, the smarter the AI is. It’s really amplified intelligence, more than agentic intelligence."

We have to put AI's work into strict boundaries: cut its tasks up in small parts, verify each of those parts, and meticulously analyze the output to improve it.

"AI doesn’t take your job, it lets you do any job. Because it allows you to be a passable UX designer, a decent SFX animator, and so on. But it doesn’t necessarily mean you can do that job well, as a specialist is often needed for polish."

The reason we have a lot of AI slop. This is why my credo is: If you can't verify it, you shouldn't use AI for it.

I'm not sure whether I agree with the next point though:
"AI is better for visuals than verbals. That is, AI is better for the frontend than the backend, and better for images/video than for text. The reason is that user interfaces and images can easily be verified by human eye, whereas huge walls of AI-generated text or code are expensive for humans to verify."

That it's easier to verify visuals doesn't mean AI is better in doing visuals. I actually think the visuals we currently see are underwhelming. Yes, generated images and videos are impressive, but not when you have requirements. And yes, I do know that it's much harder to generate visuals than text.

All in all great points!