I've found most work communication apps not to be very condusive to it, but Discord is pretty good.

Implicit management of async operations is something I hope I never have to deal with again.

Ive's work is bubbly symmetric bland crap.

SeedLM from Apple is an interesting approach for inference memory efficiency. I'd like to see someone try and build that into training so that it's not a post training compression step.

For example in XHTML you can use custom elements as table rows or cells (provided you give them the correct role and CSS display property). This is because XHTML does not modify the tree during parsing, unlike HTML which will hoist out custom element children of the table to the table's parent.

The worst of the changes is "lib.node included by default", if I'm writing Deno or web code I absolutely don't want node types included by default. Those types were a pain to deal with even in Node projects, resulting in multiple tsconfigs to avoid those types polluting platform agnostic or web code.

If Deno continues this trajectory then there is less and less reason to use it over Node.

But it's possible that all 3 Deno, Node, and Bun could share some code in the future considering they now all require Rust as part of their build process.

There's no connection to authoritative approval, other than making ecosystems without or without that kind of strict approval safer.

I should be able to limit what binaries extensions have access to though.

Part of what seemed good about Zed was that extensions have explicit permission controls.

Zed is the closest thing I've found to meet my needs, and I do plan to try it. However it's dev container support looks to be lacking in some important ways so we'll see.