Of which about 40 minutes was watching an accelerated video about Cursor's capabilities.

The goal was simple - to get hands-on with vibe coding.

I've been reading, hearing, and seeing a lot lately about developing products using AI models without a single developer.

I wanted to walk the path of a samurai vibe coder myself.

And honestly, I'm still in mild shock. Because never has launching an MVP been so fast and easy.

The idea of the experiment was as follows

1) No money
2) No development skills
3) No team
4) No product description

Uh.. Create a product in 1 day

That works and can be used

(param-pam-pam-pum)


It didn't work out completely without skills.

I can code a little, worked with VS Code and GitHub, know the words "fix a bug" and "deploy"

But objectively, that's a bit too little to do anything sensible on your own.

The other points were successfully completed.

What exactly I did

I just asked the free versions of AI models to make my personal website and web tools for pricing.

As for details, at most "I'd like a quiz, a services section, and a blog"

That's it. No logic, layouts, texts, prices - nothing.

Then the AI coded the code
(about 7-10 minutes per product)

The rest of the time was spent using the same AI to fix errors that prevented the projects from uploading to the server.

To be fair, there were plenty of bugs. The code wasn't the cleanest.

But again - 3.5 hours
For 2 MVPs from scratch

Plus, I didn't even read what they were suggesting I create, fix, and commit

Copy-pasted


Here's the website and here are the pricing tools

I'm posting them as is
Without edits, with all bugs
(say, on mobile only 1 out of 4 tools is visible)

BUT! Straight from the AI's pen
All words, prices, logic - all by it
(just so you know)

What's the point of all this

I am immensely happy.

No. Not because of my products or skills.

But because this speed of releasing creates enormous product competition

There will be many products.
Even more MVPs.

But clients with money,
who will buy these products..

The unit economics for 9 out of 10 will fall apart even faster. After all, the market remains the same size.

Now we'll have to work much more actively with value, segments, prices, monetization, acquisition, retention.

(and that's actually my thing)

Of course, you can keep saying that AI-generated code is raw, insecure, and all that.

But let's be honest..

In a year, it will definitely be at the level of a solid mid-level developer. Ready to work for you 24/7 for $20 a month.

For initial hypothesis validation, such a team is more than enough, in my opinion.


Price is more important than product. Period.

I'm off to do customer development
Which I wish you too

#value_search