There’s still work to do, things to fix, others to redo, and probably a few things I can’t even see anymore after staring at them for too long. I’m going to try to let it live now and improve it gradually. Otherwise it’ll be finished just in time for retirement, which slightly defeats the purpose of a portfolio.

Work-wise, I should soon be getting back to the project that had been postponed. And I’m actually quite happy about that: apparently it wasn’t just an elegant way of not telling me it had been abandoned.
A few other enquiries here and there too, but nothing signed yet.

So in the meantime, I’ve been experimenting quite a bit with Codex and OpenAI’s strangely generous approach to resetting usage limits. As long as it works, I’ll keep using it.

On the AI side, things are getting fairly strange.
The people building the most powerful models are starting to explain that maybe we should slow things down a little, while some researchers are leaving the ship behind them with messages that don’t exactly radiate serenity. It’s reassuring that they’re finally looking for the brake pedal. But when the people who already have a three-lap lead suddenly start asking for barriers to entry, you can also wonder whether safety is their only motivation.
Both things can obviously be true at the same time.

Another thing that caught my attention: the iPhone Duo. Not so much because it folds, but because of the shape of the screen once it’s open. We’re heading back towards something much squarer than the screens we’ve got used to.

I don’t know yet what kind of constraints this will create for interfaces, images or video, but it does look suspiciously like a return towards 4:3. After twenty years of making screens wider and wider, we’re slowly heading back towards the TV I grew up with. Full circle. PAL/SECAM, here we go again. All we need now is Teletext.