~ / Journal / 2026 / March / 6

Journal

Good night, Internet!

I spent the day vibecoding, and do not have the feeling that I have accomplished anything.

I was doubting about SaaS being potentially dead.

I am unable to vibecode a good SaaS, ready for market, from A-to-Z.

Despite me being a power-user, 2 parallel Claude Max subscriptions, Ralph Loops, not Ralph loops, good planning, no planning...

SaaS are definitely NOT dead.

I do not have any reason to doubt about it, it's all a psyop lead by slopAI leaders.

LLMs are still really dumb.

I bought yet another domain name: podcastsdatabase.com.

For some reason, I am obsessed with podcasts, and I should ship this website, starting very slowly, with a mix of AI transcription, and a human-in-the-loop.

I will try to ship a version this night.

Just a website where you can browse podcasts (for now, one), see transcript, who's speaking, ...

This weekend, I need to ship the tiniest version of Palmframe: a simple widget to get feedback on a SaaS.

Even better if I write the code myself, and get to send the messages to start the week with potential customers.

I read some feedbacks about Canny, the leader on the market, Pieter Levels complained about it, and churned to Featurebase. Some french actor complained about paying tons of money for it as well. Market is there, when companies churn, they churn to another actor, they do not rebuild it themselves.

But I should probably ignore the competition.

Let's follow DHH playbook there.

Copy the initial value prop: get feedback on your software.

Then, let's iterate with end customers in mind, directly!

I can look at websites on TrustMRR. If they do not use featurebase, or Canny, or Intercom, just ping them "hey, i'm building this, i can set this up for you in 5 minutes.

Good night, again, Internet!

I shipped... something... It ain't much, but it's honest work, as says the meme.

A prototype for Podcasts Database.