We are in the wild west of AI. No one truly has agentic workflows "down pat" yet. We have no idea which AI workflows, pipelines, architectures, and platforms are going to stick around long term. For all we know a kid in high school can be the one who comes up with a SWE workflow that FAANGs and the rest of the tech sector ends up adopting.

Recent numbers from Google suggest that Ed Zitron might very well be wrong about the data center build-out, which is the single biggest crux of his argument. But even if he's wrong on that, he'll still get his 15 minutes of "see I told you so" (while still being wrong) because even if the data center rollout doesn't implode from assets depreciating too quickly, a correction will still need to happen on the dirt end of the shovels. Even if you don't have too many shovels (which Ed thinks we do), you still need to be digging in the correct patch of ground to strike a gold vein. There's going to be winners and losers in this thing, and the market correction will happen once the biggest investors and hedge funds have correctly identified who the losers are.

What I can't do is provide a timeline. I don't claim to be a prophet. However, if Ed is correct on the hardware buildout, the sheeit will hit the fan once amortization schedules hit critical mass. Otherwise it will be whenever the big money learns who the winners and losers are on the implementation rollout. That can happen at any time. It's perfectly normal and is not a doomsday clock. That's why it's called a market *correction*.

Ed's primary gripe is that he thinks the business models aren't viable for profitability. But Google's AI infrastructure buildout is already spending less then the depreciation value of the hardware, meaning it's inevitably going to become profitable - at least for Google. Microsoft has since adopted the same approach that Google is using, focusing on faster and more efficient models in order to reduce costs.

Ed won't acknowledge that the paradigm shift for programming and SWE has already happened. He won't acknowledge that roughly 30% of radiology labs in the US and 40% of dental practices have adopted AI.

I personally think Ed is digging a hole he won't be easily able to climb out of.

> No safety disclaimers unless there is a genuine life-safety or legal risk.

> No "Note that...", "Keep in mind that...", "It's worth mentioning..." soft warnings.

> Do not create new files unless strictly necessary.

Nah bruh. Those are some terrible rules. You don't want to be doing that.

With FormKit you'd still be using a predefined deterministic flow where all the conditionals are accounted for in advance. FormKit wouldn't be able to generate a truly non-deterministic UI that changes based upon how the user answers. If you're wanting to sue your neighbor for their tree falling onto your house, an AI powered UI for a law firm could be giving you form fields on what type of tree it was, how old the tree was, hot tall the tree was, how close it was to the property line, and what room in your house the tree landed in... All generated from how the user responded to an initial "what's your grievance?" question.

I'm not advocating for or against this type of approach - I'm just pointing out a few potential use-cases for it since that's what you asked for.

https://developers.googleblog.com/introducing-a2ui-an-open-p...

In the real-world, to me it makes the most sense for client and patient onboarding.

A potential client providing information for a law-firm regarding their grievance.

A patient filling out the medical questionnaire prior to their first visit to a medical practice.

Rather than having a fully deterministic form, you'd be providing them with forms that adapt to their specific issue. The data can then be intelligently stored both as JSON and a more generic record in an RDBMS.

That's just my initial thoughts.

Google has a similar project called A2UI: https://developers.googleblog.com/introducing-a2ui-an-open-p...