I keep getting these newsletters and seeing articles about dev agent rankings and for some reason Cline is almost never there? Despite being consistently #1 on openrouter. I don't get it.

I wrote a couple of tracks on Fasttracker ~28 years (!) ago... Sadly I don't think I have the files any more but I could probably recreate one, if I could find the samples that is!

scary how quickly you get used to the quality of newer models. Perplexity's default model feels annoyingly dumb now...

It's a mistake to imagine that a type system should be judged by what it enables you to do.

A type system should be judged by two things:
* what it enables you to not do;
* what it enables you to promise.

Only took me 13 years but I finally implemented this little feature
github.com/mausch/Quart...

I'm finishing 2024 in style: trying to convince AWS support that the new web console design broke horizontal scrollbars 🤦‍♂️

Hyrum's second law: it doesn't matter how much you warn your users that your service is eventually consistent, some idiot will depend on it being strongly consistent.

Trialing copilot PR reviews:
2 PRs: 1 Python, 1 C#
It made 3 comments.
All complete nonsense. Doesn't even understand C# syntax.
At least it's more concise than CodeRabbit.
Has anyone had any positive experience with any of these AI review tools? Seems they're getting things fundamentally wrong?

Thinking about the time a senior developer told me they wouldn't learn some tech I introduced because "it didn't fit their personality"

With widely available AI services I find myself expecting senior coworkers to use them before asking for help... much like google but a much wider scope of course.
Higher productivity is great! but I'm afraid this leads to less collaboration, not sure I like that