- Vertical activity bar (https://github.com/zed-industries/zed/pull/58868#issuecommen... - rejected because it "doesn’t match our design vision for Zed")

This is my #1 annoyance: clicking microscopic icons at the very bottom of the screen to switch views.

- Don't crash my entire PC when there's a large (100+ MB) JSON file in the working directory

- Don't freeze when resuming from suspension (I suspect this is GPU-driver related on linux) forcing me to `killall zed-editor`

And it all seems arbitrary anyway. Why are trigonometric functions written in english ("tan" being short for "tangent") but other stuff uses greek letters and other stuff still uses esoteric/abstract symbols? Why is the integral symbol shaped the way it is, and why use super/sub symbols for the bounds versus `integral [0,Inf] ...`?

In my ignorance, I'm assuming math is the way that it is because that's the way it's been for centuries, and messing with it harms its ubiquity. Math notation is not the way that it is because it's particularly well thought out. It's centuries of legacy tech debt that can't be changed. Kind of like how English is a crappy language in a lot of ways, but we can't change it now because too many people use it and you'd never get enough momentum to switch.

`sum function(x) from x=0 to x=infinity`

Mathematics would be much more approachable if it just used plain English like `sum(0, Infinity, my_func)` instead of a big Greek sigma with nested function nomenclature. But on the flip side, mathematics being its own language means that a mathematician from any country can read and understand mathematics from a different country without needing to translate words such as "sum" and "infinity"

I'm using a TS type stripper API from a go package inside the server. I'm not actually doing any "bundling" which implies resolving imports, tree shaking, combining multiple files into one, minification, uglification, etc. I'm strictly stripping out types and serving what's left directly to the browser. This costs less than 1ms of cpu time on a cold load (and usually costs nothing because of front-end caching/etags)

Look, I don't want to argue with you. Just sharing my setup. I develop medical device software which means dependencies are very expensive because they incur regulatory burden. SBOMs and associated CVEs have to be tracked and reported to the FDA. My TS/golang stack means:

- My docker images can be from scratch or from busybox with a statically linked go binary

- My frontend can be vanilla TS with zero dependencies (npm or otherwise)

- Debugging is dead simple: set breakpoints directly in dev tools and it's WYSIWYG (no map files needed, etc)

- Feedback is instantaneous. If I change a file in TS and hit refresh, the change is reflected instantly (compare to bloated Vue/React shops I've seen where every change requires a 10 second frontend compile pipeline to run before you can get feedback in the browser)

1. Browser requests /js/foo.js

2. Server middleware checks if /js/foo.ts exists

3. If yes, server middleware strips types from the file before returning it

In golang I use the go package github.com/evanw/esbuild[1] to do type stripping on the fly. The middleware looks like this:

``` return esbuild.Transform(string(fileBytes), esbuild.TransformOptions{ Loader: esbuild.LoaderTS, Format: esbuild.FormatESModule, }) ```

It only takes a few microseconds and I don't need any bundlers or tsc or anything like that. Everything on my frontend is 100% vanilla TS; I make changes directly to TS and then hit refresh in my browser and the changes are reflected instantly without needing a bundler or even node/npm for that matter. Note: for this setup to work, your front end TS has to use ES modules import/export which browsers natively support. If you try to use CommonJS or something like that, you would start needing a bundler again because browsers can't resolve "require" statements.

In node it should be even easier than Go because Node added native type stripping starting with v22[2]

But even on older versions of Node, a type stripping middleware would still be very easy to implement[3][4].

[1] https://pkg.go.dev/github.com/evanw/esbuild

[2] https://nodejs.org/api/module.html#modulestriptypescripttype...

[3] https://github.com/bloomberg/ts-blank-space

[4] https://esbuild.github.io/api/#js-async

I highly recommend the introductory chapter to "Dr. Bernstein's Diabetes Solution" by the way, even for non-diabetics. It's basically just the "Life and career" section of his wiki page, but in way more detail -- a really interesting biographical account about an industrial engineer doing diabetes self-experiments with a glucose meter he procured through his wife and going up against the medical community/orthodoxy and failing, only to finally break through when he got a medical degree late in life. I could probably upload and link to just that section if people are interested.

[1] https://en.wikipedia.org/wiki/Richard_K._Bernstein

"Anthropic announces Project Disarm, a new model designed for 3d printer manufacturers to quickly infer whether the intent of an stl file is a weapon. The printer first submits the job to the cloud, and only after it's approved will it print."

Not that I want this future, just that I can imagine it.