> Most incidents resolve on their own.

This is most definitely not true.

Aaaah.

There's a station on the main line that loads full sized cars with tanks on them onto little bougies that take them up into the mountains for training.

For one time, we can be grateful that the breakdown in direct democracy is gonna save us from an own goal.

I wonder if this is a natural law, or emergent behavior of complex systems?

- a stateless dashboard for work that collects from 6 other APIs

- a refactor of a huge function with 8-deep indentation into readable small functions

- a road trip game for my kids where you take photos of things from the car

The day before I half diagnosed a cloud-init problem to the point I was tired and grumpy. I gave the AI a login on the instance showing the problem and it found it in a few minutes. All that was left was for me to write it up for the client to understand.

To me, it's a meta "oh shit" moment, where I realize I get so much small value from AI I don't want to live without it, despite the costs I'm aware of (and those I don't know about yet).

Just like a smartphone, 15 years ago.

(gestures wildly while changing lanes in his Fiat 500)

I also use Qwen 3.7 27b at work and I agree with the author: it is perfectly capable of the jobs I give it.

Priorities. Sigh.