The road to V0

Warning: You are about to read the ramblings of the world's shittiest developer.

Proceed with caution.

So, why would anyone spend 4 months (?!) trying to make a server in Python?

We thought it would be easy

By god what a mistake.

It did sound easy!

Heck, we're just supposed to call asyncio.start_server() with extra steps!

The bar is so low in Python!

And yeah, writing this wrapper was kinda fast, and I thought I was 90% there,

And as always the remaining 10% were the actual hard part.

The hard questions, the devil in the details:

• How to handle this edge case?
• Should I make everything async?
• Should I check for this or that first?
• Should I include this feature or not?
• Should this be code or docs?

The scary questions, the stuff I usually skip:

• How to measure performance?
• How to test the code?
• How to publish on Pypi?

Well...

We thought it would be cool

Let's be real, this was also a bit for (dev) street cred.

We're always dicking around in the Datastar discord about using as little code as possible (Datastar used to flex the size of "core" on the main page, as a sort of "Let me guess, you NEED more?" kinda shit).

So the initial motivation was a mix between "stripping down other frameworks to just what we need" and the infamous "not invented here" syndrome.

But where's the glitter now? Now that people shit vibecoded liquid glassfests every 15 minutes on social media by abusing ChatGPepTides?

And I look at my code and it looks like something Claude could oneshot during the time of a toilet trip.

And I look at my code and I see garbage.

We thought we would enjoy it

Some part of me wanted a long project.

Something I could come back everyday, write slowly, write a little, like the crafting of a book.

To grow through discipline, every day a minimal, fixed amount of work, in the hopes that something would happen.

And what actually happened?

Instead, I asked Claude stuff everyday.

Still I handwrote everything. Because as Richard Guindon said:

And Claude is Anthropic's way of telling you your ideas are terrific , and could really use juuuuust a few more tokens.

So I feel like I got the worst of both worlds. Like I was robbed of something.

But I can't really say "what" yet.

In the end, I learned two things:

• Yes, handwriting is important, to sharpen your sense of what is feasible or not, of how well your ideas translate to code, and to see the reasonable amount of lines you can keep in your head (500-600 for me). But let's be honest, that ships no game. Just inconsequential toy scripts that you should not use .
• The barrier in everything is procrastination. And AI feeds on that. "That project you always wanted to make, oneshot it today, don't wait!"... And that's not it. Real ballers know the solution to procrastination is more procrastination, specifically rotating procrastination, or rotproc as I like to call it. Don't want to work on your code today? Work on the docs! And if the docs bore you, advance the essays! No? Well, the website could use some improvements... Maybe add that feature? Or tweak the CSS? (I hope you like my CSS btw)... And when you're fed up with the CSS, there's always the code. Or the docs, et caetera. You won't starve if you bite more than you can chew. And you can't chew slop anyway.

Well, I got this out of my chest and V0 out of my head. Feels good man.

Now go read the code or the other essays .