~ / Journal / 2026 / February / 12

Journal

Good afternoon, Internet!

This morning, I released a few short videos (3), in the french language, showcasing Hypercode to learn PostgreSQL. I'll see how it performs, but as it was done in a very raw mode (i didn't even put effort into fix sound ^^), I guess it will only make a few views.

Anyway, I thought I'd be more motivated into promoting this platform, but not really. I feel like tons of platforms like that already exist, and I'm not sure how to really convince myself that it's worth it...

I also have to fix the Zig WASM interpreter, so it loads properly (not currently the case).

To sum what I've done the last few days:

• I created myself a "pay" button
• I launched a learning platform

In the distribution front, I think I mainly have one asset: LinkedIn.

I have a few thousand followers on X, but mostly from countries with low purchase capacity. On LinkedIn, it's mostly professional, mostly with a job, mostly (almost entirely) in tech.

When I post a low-effort post on LinkedIn, I get 27 likes. When I do on X, 1 or 2.

So, now, what can I offer to my existing LinkedIn network would be a good question.

Anthony from Cossistant had a really good idea to build an open source support platform. I learned about him on X while figuring out whether such a thing already existed. But copying, or associating myself with a guy living in Thailand, is not really in my plans. There were three interesting aspects in this project though:

• Open source (I like open source software)
• Devtool aspect (even though light)
• Almost all scaling SaaS need such a thing (whether in-house or not)

I also recalled about how much I like simple businesses like Craigslist.

So, we need to establish the following metrics, and treat this like a science:

• SNOLOC: Smallest Number Of Lines Of Code to get to buy button. Every line of code that doesn't lead to a payment is overhead. Ship the thinnest possible path from landing page to checkout.
• TTFP: Time To First Payment. From the moment I start building, how many days until someone actually pays? This is the only clock that matters.

Exercise How to apply such metrics to build my own take on Zo.computer? That is: a Linux server experiment, for AI usage.

The complexity is that they actually handle compute for their customers. I do not want to do that... Too much complexity. Any competitor with the right margins AND the right product can crush them (and I). Just like Railway with Heroku. But, knowing about Railway, you probably do not want to build such a business (I like my sleep too much).

Another subject: Discord is getting mad, and wants ID from their users, ... Maybe someone will make IRC (or something similar) the new cool?