I built a personal AI assistant nobody else will ever use. On purpose.

I was scrolling Instagram and kept seeing these clips of people with AI
assistants that launch their games, give them daily reports: "your last
video hit 4,566 views, 6.7% up from last week, keep it going. Now for
today's news..." I thought, that's awesome, I want one.

So I went to GitHub, found OpenJarvis, read the README, looked at how it
was built. And it hit me: it's good, but it's generic. It would never
actually adapt to who I am. So I decided to build MY Jarvis instead. I
named it Naos, after one of the brightest stars in the sky. Because in my
system, it's the brightest one, and I mean "system" both ways: the machine
I live in, and the orbit I move through. A star close enough to leave no
shadow on me.

Here's what it actually does for me.

I've always found the news exhausting. The same three redundant topics on
a loop, and I close the app more drained than informed. So I decided Naos
wouldn't do that. It pulls from the sources I actually care about:
hardware, games, anime. Every day it gives me a quick pitch on what's
happening in those worlds, and it can open the full article on anything
that catches my eye.

It also helps me manage my Google Calendar (full CRUD). And it knows what
games I play, what I don't play, what my friends are playing, and it
sometimes suggests I jump in and join them.

Want to steal Naos? Go ahead. But heads up, you'll need to:

• change all the Obsidian paths
• swap every API key: Claude, ElevenLabs, Spotify, Google
• redo every config file one by one, or by default it'll call you
  Nathanaël, computer science student, while IRL you're Gerald from
  accounting
• fix every broken dependency
• wipe its memory and start from scratch (so expect it to fumble for a
  while)

And that's just the start. It's unsellable. But I never had to buy it, and
I don't want it to be bought. That's exactly why I built it.

What it took: Python skills, and honestly a good dose of humility too.
Building an AI assistant that controls your PC and knows who you are is
long. I started April 28. It's June 14 and I'm still at it, still piling
on features because this project genuinely excites me. I gave life to
something unique that nobody else will ever have. If someone wants their
own Naos, sure, it'll cost them time, but they'll get an AI with their own
fingerprint on it and no one else's.

Which is where my real question comes in. On a site obsessed with "can
this become a business", I spent two months on something with zero
commercial future, that exactly one person will ever use. Was that time
wasted? Or is there a kind of value in building purely for yourself that
the "is it monetizable" reflex makes us forget? Curious where this
community lands, because I suspect it splits people.