I'm scared

A few months ago, Dario Amodei said we might not need developers in 3-6 months.

I thought that was far-fetched.

I don't anymore.

At least not directionally.

I put together a demo to show where we're at. Not because the project itself matters - it's a dumb little site that counts how many words an AI uses to describe an image (you know, "a picture is worth a thousand words").

What matters is how I built it.

I talked to Claude for ~15 minutes. Voice-to-text rambling about what I wanted. It asked a few clarifying questions. Then I told it to go build it, commit after each step, and use the browser to test until it works.

It one-shot the entire thing. React app. Deployed to GitHub Pages. Working.

In this video I also talk about: how I speak to the terminal (and it speaks back, although I fixed that aha!); how skills are incredibly powerful primitives (better than MCPs!); how we close the feedback loop with agent controlling the browser; how agents writes tests that itself understands that they would pass - all this while going from vision to deployed project in 15 minutes.

Here's what I keep thinking about:

We spent millions of dollars and years building the OpenBB workspace. Nailing the UI/UX. Speed optimizations. Infrastructure choices. Iterating on product vision.

But now? There's nothing stopping someone from opening our workspace, having Claude running on the side, and saying:

"Do not leave this session until you can replicate this product exactly 1:1."

This doesn't require our code being open source.

This doesn't even require product vision - just the ability to see the product and iterate.

And tokens are so freaking cheap.

I think we're about to see a massive consolidation phase inside large organizations. When building apps becomes this easy, every department starts shipping their own tools. Same data, different outputs. Paying 3x for the same feeds. Teams duplicating work they don't even know exists elsewhere.

That chaos creates real opportunity for platforms that unify the mess.

But!

When anyone can replicate a product by just looking at it, then the question becomes which products I want to build in-house (maintain) vs buy? It maybe starts becoming more about betting on a strong founder and team rather than the product, when deciding to buy.

You’re betting on character, authenticity and building true relationships.

Maybe the things that our digital twins won’t be able to touch for a while.

In any case, things are about to become very weird.