densh

Sqlite is essentially a user space queryable file system and it can be faster than writing to file system directly while working with small files.

As someone who spent a few years working on a compiler in this space — it’s tough, but are the results of instant startup and immediately decent performance without warm-up satisfying to use in practice. I really hope this takes off and yet another language can break free from dominant interpreter + jit compiler monoculture that we currently have for higher-level programming languages.

I usually use center 2/3 of 27'' screen with just a single top-level window for a similar reason. That puts it up to around 20.5'' and around 6:5 aspect ratio. I don't find having many windows shown at the same competing for my attention more productive. I don't benefit from having multiple code columns shown at once that much either and would rather switch among tabs or windows via shortcuts.

For me Mathematica is much more akin to numpy+sympy+matplotlib+... with absolutely crazy amount of batteries included in a single coherent package with IDE and fantastic documentation. In a way numpy ecosystem already "won" industry users over, yet Wolfram stack is still appealing to me personally for small experiments.

Coq/Lean target very different use cases.

Don't get me wrong it's a fantastic product and great price point, but the only thing it makes me think of is the complete failure of iPadOS. Ultra portable MacBook with is A18 with 8G of ram is infinitely more useful to me (for non-pen input) than full M4/M5 chip with more ram that's completely wasted due to needless OS restrictions.

I might be the only one, but it's still to this date (and dating all the way back to 2014 with the first iMac 5k display) Apple is the only company that truly gets HIDPI desktop displays with high quality gloss and 200+ ppi at screen this large. In the meantime popular and widely sold gaming screens with matte blur filters and mediocre ppi give me headache and eye fatigue after a few hours of use. Prior generation Studio Display is the only external display that truly worked for text heavy work with my eyes (including software engineering), and I'm sure the latest generation is fantastic as well.

Apart from prioritizing FFI (like Java/Scala, Erlang/Elixir), the other two easy ways to bootstrap an integration of a new obscure or relatively new programming language is to focus on RPC (ffi through network) or file input-output (parse and produce well known file formats to integrate with other tools at Bash level).

I find it very surprising that nobody tried to make something like gRPC as an interop story for a new language, with an easy way to write impure "extensions" in other languages and let your pure/formal/dependently typed language implement the rest purely through immutable message passing over gRPC boundary. Want file i/o? Implement gRPC endpoint in Go, and let your language send read/write messages to it without having to deal with antiquated and memory unsafe Posix layer.

Great summary and I think your argument is sound.

No pun intended but it's indeed an unfortunate choice of words on my part.

> As a blind person, AI has changed my life.

Something one doesn't see in news headlines. Happy to see this comment.

For example let's say I want to go to display settings from search. I enter 'monitor' in search since I forgot how it's called. First results: accessibility, privacy and security, control center, and only 4-th category Displays. It's 8-th line if you count sub-categories.

I usually google where a particular setting is now since I don't use the exact same words and the settings search is very literal.

Have any studies been done on the use of newer or less popular programming languages in the era of LLMs? I'd guess that the relatively low number of examples and the overall amount of code available publicly in a particular language means that LLM output is less likely to be good.

If the hypothesis is correct, it sets an incredibly high bar for starting a new programming language today. Not only does one need to develop compiler, runtime, libraries, and IDE support (which is a tall order by itself), but one must also provide enough data for LLMs to be trained on, or even provide a custom fine-tuned snapshot of one of the open models for the new language.

I just can't believe how dysfunctional something as basic as search over settings is now.