The LLM owners also don't like to burn the knowledge into model weights or prompts, sharing some concerns of the post, e.g. freshness. RAG is the new hotness.
Reading the original article I reflected: "I can totally see how this can happen, but for some reason it doesn't happen where I live"
I moved to Switzerland 9y ago. People care. I believe this is due to high trust society which evolved not that long ago from small, poor, tightly woven communities.
I have the book and don't speak Ruby at all. Nevertheless, is so we'll written, that you can take the lessons and apply them in any language with virtual polymorphism. A Python version of the book means bigger audience, to the benefit of the trade.
Seriously, the competing library problem is stale news. It might've contributed to things stalling a long time back, but I think the main reason right now perhaps is due to no mainstream FOSS software that is used by a large population being written in D. Nowadays, there are too many alternate choices in capable languages with more backing and some of them have more mind-share like Go and Rust and Zig. Mostly Rust perhaps. And C++ is catching up. D doesn't really have a niche where it might be a good language to use.
"D doesn't really have a niche where it might be a good language to use."
I'm wondering if metaprograming would be a niche D could exploit. I have yet to find a systems language with the same introspection and code generation capabilities (minus JAI but its non public). Though not sure what business use case that would fall into... You are right C++ is catching up, Andrei Alexandrescu has a talk that about proposing changes to constexpr to effectively recreate D's static if so he can do Design by Introspection in C++ (https://www.youtube.com/watch?v=tcyb1lpEHm0)
--Examples/more info--
They literally have a "function" that takes in a regex and spits out D code that then gets compiled in to the executable.
https://www.youtube.com/watch?v=0lo-FOeWecA DConf Online 2021 - Metaprogramming in D - Bradley Chatha (this one gives a great overview + quick dive into what D meta programing can do)
Right, but D burned some of its very valuable growth time stagnating, and probably turned a lot of people off from D. It's a nice language now, but how many people never did (and will never now) give it a chance because of the "standard" library?
When he overlaid the gray boxes over the tracks I thought “that’s a VI!”.
I’d like to point that LabVIEW is not doing any of the fancy type stuff explained here (Either, Maybe, etc). It’s just that VIs (or functions in common parlance) can have multiple inputs and outputs, and LabVIEW’s graphical approach lends itself well to this use case. But LabVIEW’s type system is fairly primitive.
I guess it's a way to optimize the build somehow, but you can achieve the same result with a proper build system, well-defined targets, and precompiled headers, but it's more work for the developers.