2026/02/04

Newest at the top

2026-02-04 12:54:12 +0100 <mesaoptimizer> went from Haskell to Rust to C IIRC
2026-02-04 12:54:11 +0100 <tomsmeding> well interaction nets is what it was all about anyway from the start
2026-02-04 12:53:58 +0100 <mesaoptimizer> he's gotten very distracted rewriting his interaction nets compiler
2026-02-04 12:53:56 +0100 <tomsmeding> name matches, in any case
2026-02-04 12:53:48 +0100vanishingideal(~vanishing@user/vanishingideal) vanishingideal
2026-02-04 12:53:41 +0100 <mesaoptimizer> https://xcancel.com/VictorTaelin/ do you mean this guy?
2026-02-04 12:53:32 +0100 <tomsmeding> but this does fancy evaluation strategies + had a hype cycle a few years back, so fits your description
2026-02-04 12:53:15 +0100qqq(~qqq@185.54.21.178) (Ping timeout: 244 seconds)
2026-02-04 12:52:58 +0100 <tomsmeding> I dunno
2026-02-04 12:52:49 +0100 <[exa]> but where's the explanation with the MAGIC gif
2026-02-04 12:52:16 +0100 <gentauro> tomsmeding: yeah, that's who I'm thinking of
2026-02-04 12:51:31 +0100 <tomsmeding> *over
2026-02-04 12:51:25 +0100 <tomsmeding> yeah graph reduction evaluation strategies for asymptotic performance improvements of standard evaluation
2026-02-04 12:51:07 +0100 <gentauro> Victor Maia is the guy doing crazy lambda calculus right?
2026-02-04 12:50:11 +0100 <tomsmeding> oh right it was renamed, I guess HVM now
2026-02-04 12:49:53 +0100 <gentauro> I know him :)
2026-02-04 12:49:44 +0100 <gentauro> «Mads Tofte (Main architect of adding region support to the MLKit, co-developer of MLKit version 1)»
2026-02-04 12:49:00 +0100 <tomsmeding> memory worked
2026-02-04 12:48:58 +0100 <tomsmeding> Formality by Maia Victor
2026-02-04 12:48:52 +0100 <[exa]> kinda looks like it disappeared from the internets, I'd say that confirms a hypothesis
2026-02-04 12:48:39 +0100trickard_trickard
2026-02-04 12:48:37 +0100 <tomsmeding> lol
2026-02-04 12:48:27 +0100 <[exa]> yes
2026-02-04 12:48:12 +0100 <gentauro> tomsmeding: yes
2026-02-04 12:48:03 +0100 <tomsmeding> I know which you mean but names
2026-02-04 12:47:54 +0100 <gentauro> xD
2026-02-04 12:47:20 +0100 <[exa]> gentauro: nah it was some weirder one, unrelated to elm, it claimed it has waaaaaay better evaluation method than stg
2026-02-04 12:46:44 +0100 <mesaoptimizer> tomsmeding: yes. https://github.com/melsman/mlkit
2026-02-04 12:46:25 +0100 <mesaoptimizer> I think the MLkit links are broken but IIRC they still have a github
2026-02-04 12:46:16 +0100 <tomsmeding> gentauro: do you mean this? https://elsman.com/mlkit/
2026-02-04 12:46:10 +0100 <mesaoptimizer> gentauro: MLton and Poly/ML are rock solid though
2026-02-04 12:46:05 +0100 <gentauro> https://www.roc-lang.org/ <- this link instead
2026-02-04 12:45:23 +0100 <gentauro> not to confuse with https://en.wikipedia.org/wiki/Rocq (rename of Coq)
2026-02-04 12:44:59 +0100 <gentauro> https://www.roc-lang.org/examples/ElmWebApp/README <- this one?
2026-02-04 12:44:17 +0100 <gentauro> the one that spawned from Elm?
2026-02-04 12:43:29 +0100_JusSx_(~jussx@37.161.247.165)
2026-02-04 12:42:18 +0100trickard_(~trickard@cpe-61-98-47-163.wireline.com.au)
2026-02-04 12:42:05 +0100trickard(~trickard@cpe-61-98-47-163.wireline.com.au) (Read error: Connection reset by peer)
2026-02-04 12:41:41 +0100 <[exa]> with extra LLM flavor on top
2026-02-04 12:41:34 +0100_JusSx_(~jussx@78.210.76.107) (Ping timeout: 256 seconds)
2026-02-04 12:41:33 +0100 <[exa]> seems a bit too much like the "works by MAGIC!!!!" 100% better FP compiler that we had here like 1 year ago (what was the name btw)
2026-02-04 12:38:46 +0100 <gentauro> seems too much
2026-02-04 12:38:41 +0100 <gentauro> so this core components are mental. Then they are also opting for creating all the environment tools https://speakez.tech/blog/leveling-up-with-lattice/
2026-02-04 12:37:29 +0100 <gentauro> The F* "thingy" is a 100% copied from Liquid Haskell -> https://software.imdea.org/careers/2022-02-intern-liquid/
2026-02-04 12:36:48 +0100 <gentauro> [exa]: MLKit comes from academia -> https://www.itu.dk/research/mlkit (and the reserach page doesn't even work, so yeah, 100% academia xD)
2026-02-04 12:34:59 +0100fp1(~Thunderbi@2001:708:150:10::9d7e) (Client Quit)
2026-02-04 12:34:30 +0100fp(~Thunderbi@130.233.70.158) fp
2026-02-04 12:33:32 +0100 <[exa]> wait how do they get knowledge from academia? (I probably missed that part, looks super industrial to me tbh)
2026-02-04 12:32:31 +0100 <gentauro> but, that sometimes come with the downside of not being "industry friendly" as you mention
2026-02-04 12:31:59 +0100 <gentauro> [exa]: yeah, it sounds to good to be true. But I like how they go to academia to get knowledge.