davvid

This is nothing new, tho. The downfalls of reward maximization has been a known issue without a solution ever since reinforcement learning was first researched.. in the 1980s.

No, Git is alive and well. Git is under active development and continual iterative improvement.

Github might be on the decline, but git != github

git-difftool is quite helpful, especially if you want an easy way to integrate existing file-based diff tools.

Nor Catmull, unfortunately. FWIW, the article is centered around the financing and IPO side of the story.

> We did not want to spend time maintaining a backward compatible parser or doing code archaeology. So this option was discarded.

Considering all of the effort and hoop-jumping involved in the route that was chosen, perhaps this decision might be worth revisiting.

In hindsight, maintaining a parser might be easier and more maintainable when compared to the current problems that were overcome and the future problems that will arise if/when the systemd libraries decide to change their C API interfaces.

One benefit of a freestanding parser is that it could be made into a reusable library that others can use and help maintain.

    > I had to get my project to emit compile_commands.json,
    > get clangd, figure out which things about our build
    > process clangd was not understanding from 
    > compile_commands.json and add them in .clangd

That sounds rough. This is anecdotal, but in my Linux corner of the world, ccls has been an easier user experience as a C and C++ LSP since I've never had to resort to messing with flags in the generated file.

I haven't used clangd myself, though, so I can't say either way, I just know ccls works well.

By convention I tend to have the generated build system in build/ at the top-level of the repo so that the file is at build/compile_commands.json. That, or I'll arrange to have a symlink there pointing to one generated elsewhere.

The nvim snippet I use in my init.lua to setup ccls to work in that scenario is:

    vim.lsp.config('ccls', {
        init_options = {
            compilationDatabaseDirectory = 'build',
        },
    })
    vim.lsp.enable('ccls')

My actual config does also contain a capabilities = ... argument that forwards the default_capabilities() from nvim-cmp, but you get the point. I hope that helps in case you're curious to give neovim another spin.

> upon further thought, I don't like having to move my hand to the arrow keys

ctrl-n and ctrl-p should let you select next and previous from the autocomplete menu while avoiding the arrow keys.

> Throughout all of this, Horse News was the only stable work I had.

Only a good writer, that truly enjoys their craft, is able to masterfully insert a witty dry pun like that into their work. Bravo!

The article didn't mention Python's atomic refcounts. They're very bad for cache usage as they're constantly invalidating cache.