On X
Posts about software and AI: building products, programming and the AI tools I use.
159 posts
The problem with AI is that it makes solving problems so easy that we try to solve them all. Even the ones that don’t matter.
The important thing is to focus on the bottleneck.
There’s no point solving 20 minor problems if you don’t solve the one big one.
Which private company would you like to be able to invest in?
Here are mine: → 🚀 SpaceX → 🤖 OpenAI → 💰 Revolut → 💸 Stripe → 📊 Bloomberg → 👾 Valve → 🎮 Epic Games → 🛒 Mercadona
You’re gonna laugh, but I was waiting for GPT-5.6 to handle three different bureaucratic processes.
Technology saves our asses again.
If you have a Mac, I assume you use Raycast
If not, you don’t know what you’re missing
One hugely underrated thing about OpenAI is their taste in interface design.
They’re beautiful, easy to use, intuitive, and feel good to use.
xAi is surprisingly pretty decent too.
Google and Anthropic, on the other hand, are the exact opposite: interfaces that make you want to rip your eyes out
On the other hand, whatever AI takes away, AI gives back.
Fix your landing page in a couple of minutes
🇨🇭 The lie I was told before moving to Switzerland: "You need to speak German or French to work here".
False.
In tech, English is enough. If you're good, of course.
We’re reaching a curious point.
It’s not that I don’t code anymore. I don’t even write.
I just tell the computer what I want it to do… and the damn thing figures it out. Builds it. Works.
Given OpenAI’s historical tendency to always want a better model than Anthropic*, GPT 5.6 can’t be far off. In fact, it could be this Thursday.
I’m hoping for a model in line with the previous one: a bit smarter and a bit more token-efficient.
*The sad thing is that GPT 5.5 is still better than Opus 4.8.
What pisses me off most about ElevenLabs and AI voice is that young people will never know Loquendo
