Sure sounds like an LLM.

It’s not exactly like vim, and there are plenty of vim plugins that don’t work with it (anything vim8 onward). There has never been a 1-to-1 correspondence, the gulf widens as both develop different features with different philosophies.

The most egregious offense on Neovim’s part that I can’t get past is the removal of access to the shell in which you run vim (via :!, :w !, etc.). Vim is so much more capable of being closely intertwined with the shell, whereas neovim requires everything to be done through terminal buffers (speaking of which, vim’s terminal buffers are a lot better than Neovim’s).

Also, Lua is really overrated and worse for vim scripting than vim9script (which is both more native to vim and faster).

Most of vim is not emulated. It’s very surface-level and limited. The closest is evil mode for emacs, which is decent, but still lacks a fair bit. The emulators in Intellij and VsCode are paltry in comparison to what vim can do.

That’s why vim is so great: it has a ton of power built right into it without customizations, and it’s already installed on basically any unix-like system. Unlike, say, vscode, it can do a ton of stuff out of the box without any plugins at all.

It’s a tool with a medium-high skill floor and incredibly high skill ceiling. It rewards investment and is something that is able to accommodate one’s growth in skills rather than holding them back with limitations like typical editors do. Its built-in scripting is a big part of that and is something that really sets it apart from editors like vscode. And it’s much, much faster and lighter weight/less memory-intensive than other editors.

It’s not as big of a deal as you might think. You still have a lot of your muscle memory from regular keyboards. It might take a little while to adjust when switching between the two, but it’s not that bad.

If you switch between the two enough, you can actually type on both equally well.

A lot of mechanical keyboards these days are programmable using QMK Firmware. I actually use https://www.caniusevia.com/ instead though, which uses (a subset of) QMK under the hood but allows programming the keyboard via a Web app on the fly.

For my layout, I have the standard QWERTY layout for the unmodified layer (layer 0, holding no keys). Then I can hold down a thumb key for switching to a different layer, which has things like symbols, F1-F12, Home, End, etc. The layout I use isn’t too far off the default Iris layout, just a few tweaks here and there (like one that allows me to hold a key for control, or tap that key for escape).

Vim has an entire dedicated scripting language built right into the editor and accessible while editing.

Even without plugins, sometimes certain things can be too slow and you want to stop them.

Ctrl-C absolutely should not exit. There’s plenty of times you want it in vim to interrupt something in the editor.

As others have said, it’s on the screen if you open vim without a file. Otherwise, it’s a tool for people that bother to learn how to use it. As someone who has been using it daily for the last 10 years, I would find it incredibly obnoxious to have a bunch of useless screen clutter telling me basic things that are easily learned.

It’s simply muscle memory. You think of the action and your fingers do it faster than you can consciously think of where they need to go. But I also use a split ergonomic keyboard (the Iris) and have symbols accessible from home row behind a layer. Though I can switch to a standard keyboard as needed too.

Been using it for all of my software development for the last 10 years. It’s fantastic.

That’s what happens when you start using LLMs for all of your software development. Garbage code all day long.

Because they vibe code the shit out of everything now. Insane shit is bound to happen.

Oh for sure, it’s definitely less common, though there are a number of other companies running it in production as well. This isn’t my first Haskell job, after all. My last job was also a similar size company and codebase. Facebook was even running it for a while for their sizeable abusive content detection system before that was shuttered due to company politics/policies (back when they were trying to do something about it at all).

But yeah, it’s not the first pick for a lot of companies, though I tend to think of that as a simple mindshare/inertia explanation than anything inherent to the language/ecosystem.

I’m not entirely sure what the Haskell comment is supposed to mean. Just that it would be cool? Or that it would be hard?

Because if it’s the latter, that’s just really not true. Haskell’s quite well-suited to writing web servers and has many high-quality libraries for it. I’d actually argue that it’s one of the best options for it in 2025. I write Haskell professionally developing web servers we use for our web and mobile apps. We have about 0.5 million lines of Haskell in production (and given how terse Haskell is, you could expect that size to be at least double were it written in an imperative language).

Heh, guilty. Pretty much always something where I had an ambition to make a change but got distracted or didn’t have time to work on it.

They aren’t useful now, but even assuming they were, the fundamental issue is that it’s extremely expensive to train and run them, and there is no current inkling of a business model where they actually make sense, financially. You would need to charge far more than what people could actually afford to pay to make them anywhere near profitable. Every AI company is burning through cash at an insane rate. When the bubble pops and the money runs out, no one will want to train and host them anymore for commercial purposes.

LLMs do not understand anything. There is no semantic understanding whatsoever. It is merely stochastic generation of tokens according to a probability distribution derived from linguistic correlations in its training data.

Also, it is incredibly common for engineers at businesses to have their engineers write code to automate away boilerplate and otherwise inefficient processes. Nowhere did I say that automation must always be done via open source tooling (though that is certainly preferable when possible, of course).

What do you think people and businesses were doing before all of this LLM insanity? Exactly what I’m describing. It’s hardly novel or even interesting.

All of that can be automated with tools built for the task. None of this is actually that hard to solve at all. We should automate away pain points instead of boiling the world in the hopes that a linguistic, stochastic model can just so happen to accurately predictively generate the tokens you want in order to save a few fucking hours.

The hubris around this whole topic is astounding to me.

…regular coding, again. We’ve been doing this for decades now and this LLM bullshit is wholely unnecessary and extremely detrimental.

The AI bubble will pop. Shit will get even more expensive or nonexistent (as these companies go bust, because they are ludicrously unprofitable), because the endless supply of speculative and circular investments will dry up, much like the dotcom crash.

It’s such an incredibly stupid thing to not only bet on, but to become dependent on to function. Absolute lunacy.