kylereeve

no clout

(keep on disk, don't commit)

There's a very short list of people who can get away with "Classic Torvalds"

Reminds me of Ace Rothstein wanting the exact same amount of blueberries in each muffin.

That's how all application-specific specializations work though, take advantage of domain properties that make you need a less generic algorithm.

I've never seen an IDE "mandated", I've seen officially supported development setups where you're on your own if you do anything different. Is that not the standard?

When this bubble finally pops, someone is going to have to clean up all the nonsense AI code out there.

they might be

I'm gonna be pedantic, shouldn't it be "bottom-up" instead of "bottoms-up"?

> #2 Software projects that somehow are 100% human developed will not be competitive with AI assisted or written projects. The only room for debate on that is an apocalypse level scenario where humans fail to continue producing semiconductors or electricity.

??

"AI" code generators are still mostly overhyped nonsense that generate incorrect code all the time.

> If one just chooses a reasonable documentclass and if need be a few packages suited to the requirements of one's document, then it all "just works" with (mostly) sensible defaults and minimal configuration.

Ironically, very similar to the story with modern C++. If you use a limited subset it can "just work" but only if you are disciplined and don't have to mix in legacy code that's pre-C++11.

> Does an intern cost $20/month? Because that’s what Cursor.ai costs.

This stuck out to me. How long will it continue to be so cheap? I would assume some of the low cost is subsidized by VC money which will dry up eventually. Am I wrong here?

Assuming you're not being facetious, one of the best parts about PBT is it gets you a good percent of the value of formal proof with a lot less work. PBT at least lets you demonstrate that property is ~probably~ true, whereas traditional unit testing doesn't usually explicitly state properties.

Do you have CMake actually run `git clone`, or do you clone separately and point CMake at the `FIND_X` files?

On the flip side, when you have a manager who's genuinely on your side and wants to help you produce value (seems rare, I've been lucky enough to land one or two), "pleasing your manager" can accelerate career advancement and actually delivering working software you're proud of.