Rochus

Cool. It's still a great backend, even for a statically typed language, and the FFI and debugger features are very useful. Mono is yet another great backend which in addition supports multiple th…

And a related paper by the same group: https://news.ycombinator.com/item?id=48262160

Here is the repository with the benchmarks: https://github.com/rochus-keller/are-we-fast-yet/ And here is a repository with a compiler backend to generate LuaJIT bytecode an…

This is incredible. Wouldn't have thought that it works so well in the Firefox on my old Lenovo T480.

Interesting results. Particularly they barely found a speed-up of newer compared to older LuaJIT versions (rather the contrary). Maybe they should have used the Are-we-fast-yet suite instead of the (r…

see https://news.ycombinator.com/item?id=48247125

I asked a question and it told me that I am at position 96 on the queue and can expect the answer in two hours ;-)

Cool idea. I guess a similar approach with a bit of tweaking would also work with Faust, or recently Cmajor, in combination with an LLM.

That's fascinating. This sounds like a variational autoencoder. The embeddings, which from my humble point of view (as a trained musician) are a largely unexplored field not really supported by e…

Cool. Generation of symbolic music using transformers is indeed a pretty neglected field. I assume you have to encode more "musical knowledge" into the embeddings than when you "just&qu…

> I'm starting to think Java is actually the ideal LLM language Why?

It even preceded "software".

see https://news.ycombinator.com/item?id=48157360

I always wondered why "the entire software industry" is a failure. Now we know it was because variables in Smalltalk are not first-class objects (plus some other essential issues). The comme…

I'm part of the old guard too, and I like the process itself, but I also like the result. I just think his arguments why to avoid AI sound like they're taken from an academic brochure, not l…

Would have hoped for better arguments. I got the impression that he is not actually sure why to avoid AI, but welcomes every argument without really having a concept.

Cool. What is the core benefit compared to e.g. Haxe?

That's amazing. Are there measurements of decently large clojure applications both with the interpreter and the original clojure engine on the same machine?

Amazing work, thanks.

That's cool. I wonder how you manage to keep track of everything with this huge all-in-one main.c. Do you plan to achieve 100%? Is there a list of what is actually missing and which relevant proj…

There are reports by independend testers (e.g. https://briancallahan.net/blog/20211010.html or https://news.ycombinator.com/item?id=40350384 ) suggesting that ran…

Interesting, thanks for posting. I had a first look and think that the AI has messed up with the floating-point and casting instructions (i.e. just copied over the RV instructions). I also spotted oth…

> it's like one Claude prompt to add MIPS support to this Can you provide the code somewhere? Is it complete? Does it really work? I have a hard time to believe that an LLM really can genera…

Rename it to "Pickypedia".

Interesting. How does it compare to e.g. Faust or the SuperCollider language? What's the "unique selling point" of Audion compared to other "music programming languages"?

Interesting. Reminds me of Typst (both implemented in Rust and replacing TeX to some degree) and Microtex.

Neither the whole L4 line of work, which fundamentally rely on synchronous, rendezvous-style IPC as their primary kernel-level communication primitive; the transition to synchronous IPC together with …