Working on a new SaaS prototype this weekend and trying TanStack Start.

Being in beta makes it sometimes hard to know what's the up-to-date thing to do and even harder with Cursor.

But now that I've figured everything, I think I like it more than Next.js.

Less magic, more structure.

It happened to me more the last couple of times I had jetlag than usual too.

There are some good tips here (not jetlag related but still): www.hubermanlab.com/newsletter/t...

I also saw epsom salt baths recommended but didn't do much for me.

I use LLMs in french when I want info about something french but never tried deep conversations.

Because I write things about Japan in English, I'd love if LLMs "searched" the info in their Japanese training data but I'm guessing it doesn't work like that.

Just finished migrating Japanetic to my new design.

Also made some improvements to the pSEO structure. I'm curious to see if it changes anything.

Moved from next.js to astro (the other websites were already on astro).

And removed a couple of unused features.

Alternative: Wishing side projects make enough money to stop freelancing.

Already shipping in 2025.

I've revamping all my websites. New design, new copy. The goal is to share the design across them to ship changes faster.

Still WIP.

I could have made a 2024 post gloating about my business having its best year.

But the only thing that really mattered to me was finally healing my brain fog after 10 years of trying so many cures.

I'm simplifying my SaaS.

And to not make excuses about not launching, I've coded the billing, team management, and plan limits.

Now that it's done, it's mostly just making prompts... Which is actually the hardest part.

I just reached 1,000 unique sessions per month with my programmatic and mostly AI side project, Japanetic. 🎉

Don't have much time to work on it but I could probably 2x this traffic with a couple of days of work.

Good idea. Interested to see what you come up with.

I always say that then start working on a tiny product and it becomes huge and I give up.😅

I feel good when I delete 10 subscribers. Deleting 15k is the dream.😂

Why haven't I released my SaaS yet?

I started it as something I'd use and my basic needs were just too complex.

My solution:

• Make a few scripts that take care of my needs
• Then I can focus on making a SaaS I can actually release in a few weeks instead of 6 months

It's the basic rules of messaging. It applies everywhere.

The pitfall is exactly that. I can quickly spot in a conversation if they are open to listening about my model or not. If they can't, I'll spend more asking questions that will put holes in their models.

Building in a mature category makes it easier. Yes there's still the possibility of pivoting the niche but you can write content on use cases that are shared among the different niches of the category.

Great, thx for the tip! Trying right away. Was really unhappy with my Discover feed (I don't follow enough people also).

This guy maths!

I really dislike the current Windows icons. But I'm on macOS so I don't care that much.

You launched too late if you removed the limits for yourself.

Yeah I really don't like these hustle gurus spitting on 9-5.

When I talk to wannabe entrepreneurs, they lack the skills to make something profitable quickly (even a service business). Better taking t slow.

Also a lot of these gurus get rich with scams (brainrot content, dropshipping, etc).

Wait aren't you in Portugal? I don't portray it as cold and dark, at least compared to my north of France.

I hesitated spending the winter in a sunny area and decided not to but the urge is getting stronger by the day.

You like playing with fire.😅

Don't you need $42,000/mo for the Twitter API though?

Built a Facebook page manager (co-founded + raised money) and FB changed how the pages work and it killed the app.

A few years later, got hired by a startup to make a IG analytics platform. Then FB killed IG API. I warned them though.

Happened to me 15 years ago. I've warned so many founders about it. They still built on top of big cos APIs. And they got burned.

Use social proof (results, customer quotes, etc) as close to the recipient as possible.

Otherwise, keep it short, no links (reduce deliverability), no hard sell in the 1st email.

Was talking to someone who thought outreach was dead.

I said it's because he's doing it wrong.
He said he wrote good emails.
He read them to me.
They were shit.

It's always the same story for every "dead" acquisition channel.

I can have something functional quickly but the gap between functional and releasable is demotivating.

But once you find a way to make it work, it's a great acquisition channel.

I use Sanity and I kinda like it. Periodically I go over a lot of different CMS to see what changed. I wanted to try Payload last time but 3.0 wasn't out and 2.x didn't seem great. Maybe it's a good alternative to Sanity now.

12 half-baked products with 0 customers. Nice.