Looks rad, I'd love one just like any red-blooded American contributing to the radical Biden biomanufacturing agenda

@hankgreen.bsky.social I think this would make for a great explainer video

4) They walk back the policy, and everyone thinks they beat the insurance companies, when in reality we were conscripted in the ASA's fight with them to continue getting paiiidddd.

Seems bad!

Wow. So the timeline here is
1) Anthem announces they'll pay anesthesiologists based on Medicare rates last month.
2) A vigilante murders the CEO of the most corrupt insurer this week.
3) That news stirs up anti-insurance vibes, and the zeitgeist latches on the ASA's press release on this.

Thanks, this was really interesting. Doesn't it still lead back to shifting costs to patients though, i.e. if the insurer caps how much they'll cover for anesthesiologists, won't patients need to pay the overage?

Also, appreciate the conversation! Like I said, I'm coming around to static metadata, but I think it's clear there are real tradeoffs here, and not just in switching costs.

Given that this is only a problem for local installs, allowing the readme metadata to be dynamic seems like a strict win, in that it allows developers to generate more accurate/useful readmes, but only if they sign up for that complexity. Local install of 3rd party packages is super uncommon, right?

Yeah, that's neat! But what you're now describing is a problem with how packages of different varieties (wheel, sdist, or local) store their metadata, not that it is dynamic.

Yeah, I wouldn't argue it's at all essential. I use it today in my open-source and industry packages to automate aspects of the release process, but I'm sure we could find alternative approaches.

That said, it feels a bit odd to privilege a VSCode plugin over a build plugin. In your example, that feels more like a tradeoff in functionality than a clear-cut "tiny tax". Right?

Thanks for engaging with the substance, appreciate it! I'm coming around to the tradeoffs of static metadata, but ultimately, different tools will need to be developed to solve the problems tools like setuptools-scm solve if the community goes down that path.

Oops, the last day was apparently a blip, but you get the idea: pypistats.org/packages/set...

Appreciated the substance here, but it really detracts from your argument when you sub-tweet (skeet?) projects like hatch-fancy-pypi-readme and setuptools-scm. The latter has >3M daily downloads, so it's not like it's some zany side-project.

Ah, that makes sense, but also, disappointing! Apparently Louisiana too, but they call them parishes instead of boroughs. Language is hard!