วีดีโอ

Odin ignited my love for programming just before i was about to quit, thank you Bill!

Incredible! 👏👏👏

I've listened to a bit over half an hour and what's described sounds like an absolute mess of half a dozen partially overlapping language dialects and projects. I'm not convinced this state of play is actually positive for OCaml.

Thank you both very much for this excellent video. I like Zig and will give it a spin. Having written millions of lines of code in my life, half of it in C++, the rest in a few dozen other languages, you name it, and there is a 90% chance I used it. The problem with getting old and having seen a lot is: The more experience you have, the lower the chance to find a "better" language than the "best" language you already found, leaving aside the effort to master a new language, which on the other hand gets easier the more you have seen. "Best" is of course also always specific to an environment, but as a general purpose language, C++ would be my choice if I could only have one programming language. It basically includes C, but I can have C-like control combined with metaprogramming and a ton of other perks thrown in, like the STL, the whole type system, classes, operators, overloads, variadic templates, exceptions, contracts and dozens of other things in the language. C++ is a mighty sword, but pulling it out and wielding it requires effort and can be painful, especially if you really want to unleash it's power. And the question is: Does it have to be that way? Even I, as a veteran, find myself fighting the compiler, especially when trying to use advanced template metaprogramming, with compiler error messages that are not always helpful, and sometimes stuff does not work that should work according to the standard, and then you need to nudge the compiler to do what you always as if it were an AI. On the other hand, I don't have enough control. Recently I started to write a library that allows me to write reversible porgrams in C++ that can run forward and backward, and while I can make it kind of work, I am not happy with the solution so far. The problem is a lack of introspection at the statement level, causing me to have to wrap every statement in a lambda and make my own functional control structures for loops and branches. Maybe it is an abuse of C++ what I am trying there, but I think it might have a few interesting applications. First of all, fully reversible programming like offered by reversible programming languages is kind of depresssing. What you really want is to only run parts of your code as reversible code. And I want the reversible code to be fast and not produce as little garbage as possible. Garbage is the data you have to produce to be able to reverse all of your code and is unavoidable if you want to do interesting things. From what I heard about Zig so far sounds amazing, and I think I will try to abuse it, too. I really want to see what I can do with these compile time programs.

Damn…your intros are consistently excellent. I have recently discovered this channel and watched maybe 10 interviews so far and everyone starts off with a fantastic introduction to really set the tone for the listener. 👍🏻

the thing with my adhd is... sometimes i will be extremely focused and productive, it can be one of those moments of genius that i'll do months of work in a day or maybe a few, and as a manager if you manage to detect that and avoid distracting me with stupid meetings or start micromanaging With time as a person with ADHD i'm learning to detect sooner when i'm struggling so i can communicate with the manager but here's the thing, sometimes i just can't tell, so the difficult part about managing people, specially with adhd is to notice those moments as soon as possible to maximize how i work and not just have a person that sometimes is good and sometimes bad

Nice interview, I'll probably never use zig in my lifetime. It's too complicated

Isn’t Netflix already plural, like Twix?

Was he serious about creating embeddings from the Bible?

Always able to find people doing facsinating things with tech, was really interesting. A type of dev work I didn't even realise was a thing, but makes obvious sense why when you hear it does.

Excellent interview!

I find all this stuff fascinating. Does anyone have any tips/resources to learn more about this stuff?

I pronounce the emu in qemu like the bird because the bird is part of the Logo

Another interesting one 👍

Go Evan! Love to hear he is still working on Elm, it's such a fantastic language. The error messages, the simplicity, it really is a thing of beauty.

wow!! perhaps with AI Agent he will take this to another level!!!

I would like to know why is still so popular, is there no other alternative to it? I have tried to avoid it like the plague, and only once I had to work in a place that used it but just to execute basic commands. Perhaps I haven't worked on something so complex that required it 😅

Thanks!

Really liked that (podcast/interview/whatever)!

First!

With the audio programming, if you want to target PC, there is a very good chance you’ll be looking at the VST SDK. It’s very very C++. I think there have been moves recently to be a bit friendlier to C to allow other languages like Rust to get in there, but if you look at the code, it screams C++, it’s a pointer piñata. The only newer language I have any hope for in this sphere is Zig.

Unison's unique features offer promising possibilities for enhancing the FunSearch operational model: Content-addressed code: Unison's hash-based identification of code could streamline FunSearch's program generation and evaluation process. Each generated program would have a unique hash, simplifying tracking and caching of results. Immutable codebase: Unison's immutable data structure approach to codebases could enhance FunSearch's evolutionary process. It would allow for efficient storage and retrieval of program versions without conflicts. No builds: Unison's ability to parse and typecheck definitions once, storing results in a cache, could significantly speed up FunSearch's evaluation phase. This would reduce overhead in assessing generated programs. Easy distributed computing: Unison's content-addressed nature facilitates distributed computation, which could enhance FunSearch's parallelization capabilities. This could allow for more efficient scaling of the search process across multiple machines. First-class documentation: Unison's approach to documentation as executable code could improve FunSearch's ability to generate and evaluate self-documenting programs, potentially leading to more interpretable solutions. Strong typing with inference: Unison's type inference could help FunSearch generate type-safe programs more efficiently, potentially reducing the number of invalid programs generated. These features could potentially make Unison an excellent choice for implementing core components of the FunSearch system, particularly in program generation, evaluation, and distributed processing.

Thank you very much for the very informative discussion about a great language. I really hope that somebody would make a killer app in Pony and it will attract more attention

Ok, they got there

At least the first minutes seems as they are taking about the shell not the terminal

So…learn Rust and Go and Svelte got it.

Great channel - love the diversity!

Unison hashes can be directly used by LLMs as tokens. Correct? Doesn't this have enormous implications for computational analysis?

My only complaint would be that he's underselling how exhausting it can be to have ADHD, especially if you're undiagnosed because you're "too smart" and used unsustainable coping strategies because for your entire childhood you've always been told "you're smart, now if only you put in a little bit more effort and actually used that" while giving your 100% and still falling short. It took me almost 40 years to get a diagnosis and apparently all my lifetime of mental health issues stem from my ADHD.

Just a note for the subtitles is not "RiskScript" is "ReScript" and "Melange" not "milange"

no, it's a bad idea to use python and fix it. Better use a different language with a less broken syntax and ideas.

bevy has a really unique position to be able to not even have to build an editor by being able to create a live reload ad editing experience as a blender plugin

Robert Virding’s First Rule of Programming, which states that “Any sufficiently complicated concurrent program in another language contains an ad hoc informally-specified bug-ridden slow implementation of half of Erlang.”

This was fascinating!

Have not watched this yet but I have to vent my frustration just at the intro. "has open source always been bad? this person is not getting paid enough for it" is an incredible sentiment that I want to never hear again from anyone in my life. It is open source. Can we maybe not bring weird capitalist production rules into the last space we have left? yes devs need to eat. Thats why theyre usually employed. Wtf

I would love to have tried Elm as I'd never really used a functional thoroughbred before. By the time I became aware of it, politics had apparently ruined the project :( I've heard good things about Elixr but I love my types!!

I love this episode! 🎉

That was really cool! You should do more examples.

Enjoyed listening to this for 47 minutes and not learning anything

15:57 we call that drizzle-orm, zod, trpc in typescript world

This is by far the best open-srouce channel content I have seeing in a long time, keep it up the good work💗

I really appreciate the variety in guests and the fact that a lot of them are doing weird/unusual stuff with code - music composition, language for CAD, etc. Gives you insight into stuff you normally wouldn't hear about!

"How often do you see a 30 year old codebase get a new lease of? life, it doesn't happen" - this is wrong it does happen how old are the Linux commands and the code behind them, how old is bash, and Bourne shell. They still have terminal emulator protocol code in them I believe. Sure they may not be changing (much) grep kind of changes, id imagine they have been reimplemented a lot. Point is good tools stick around, if it's useful it continues. Now it doesn't mean there aren't better ones to be discovered.

zed is a top-notch editor. I have been using it in Vim mode not thinking of going back. I come from: neovide neovim helix vs code

Hey! It’s great to meet you!

I enjoyed listening to this, my own takeup of bevy has been slow , just frustrating general gargles of an early pre v1 release. I think 10 year wait for tripple A studios takeup is a bit long. I think the opportunity for so many devs to make money while they develop games is a big one for studio pipeline productivity & collab software creation. The Modularity nature of Bevy, means that people could create their own plugin projects and charge for pro features or just have a external library or gui managing alot of stuff they sell for lots of money. While bevy community management does it's long term thing to have official systems added as standard. The main thing that's needed is to standardize the key pieces and get into a 1.0 release. So people can add on confidently knowing they arnt going to have to rewrite there entire plug-in or system every 5 months. Then you'll probably reduce that triple A takeup from 19 years to 4 or less. A gui might just mitigate it all to 3 years. Cos that will get more potencial contributors playing with bevy theirfor solving PR's quicker. Also a secondary community manager to monetise contributions is key, via an asset store early, is a huge driver to get more people on board from both a contributions and user base , it's the one thing unity got right and a lot of those assets became official in engine features because they were so well defined , needed and the creators got so much value back contributing while making their games...win win. Is just how it's motivated and initial tooling and standards are put in place that could get a model like that working for the engine growth.

SonicPi is an amazing project/tool. Thank you for this interview.

41:42 you're highly parallelized yourself 😂 What a great discussion! I find bevy quite interesting and have played around with it while learning rust. Very cool because of ecs. This concept feels quite natural to me. But I got a bit frustrated, because I'm a graphics n00b and have no idea what I'm doing. And when the screen is blank you think "At least it's blank at 120Hz" ;-) I hope this project succeeds for a long time. The community is super helpful and friendly. And after hearing about Alice's management style I think she plays a big role into bevy's success. Thanks a lot for keeping the project going!

I've really enjoyed listening to you both