File: 1759812101634920.jpg (102 KB, 630x1200)

102 KB JPG

This is an AI board luddite boy.

>>108948846
Fuck off nocoder

I wached the whole thing and the guy is more /g/ than /g/

develops with nothing more than linux+vim
uncompromising on open source tech
etc etc

im an LLM scriptbabby but this makes me want to learn zig

>>108948896
Cry harder snailcat

File: 1411310300225.png (56 KB, 401x372)

56 KB PNG

>1week, 13days, 11hours of video
Fuck off!

File: ..png (235 KB, 411x694)

235 KB PNG

>>108948771
>we banned AI because there is not enough zig code on the internet for models to be trained on and become good at zig

>>108948954
Contributing to zig is not about being "good at zig" it's about being good at constructs that make up zig. Aibabbies like you will never understand this distinction.

>>108948771
>Dead snailcat language that is losing relevance makes a self congratulatory video on why that's a good thing
Lmao. This is like some Chud making a 14 hour long video on why winning the "Most unfuckable incel" award is a good thing, actually.

>>108948950
Name one thing you’ve made with ai

>>108949107
Ok
I made this
https://files.catbox.moe/52l73w.mp4

File: 1780139387714889.jpg (229 KB, 1000x1380)

229 KB JPG

>>108948846
LLMs are the equivalent of pic related.
A literally comically inefficient technology that most people won't know existed in 50 years.

If you don't like this example, think of Space Shuttles. They worked, but they sucked. So now we're back to rockets. The same will happen to this fiasco. After all, your car will never drive you until true AGI is achieved.

>>108949151
>"i plagiarized the lowest form of wit"

>>108949187
OK, ok, here's a banger I worked on.
https://files.catbox.moe/fbvau5.mp4

>>108948950
>month old forced corporate "meme"

>>108949181
>LLMs are bad because I say so
>Also ignore all the other types of transformer architecture neural networks

>we banned AI

This is probably the shortest way someone can let others know they don't understand how to use AI.

The title seems clickbait. In short Zig got a lot of slop pull requests and it was easier to enforce full ban on AI than to define some acceptable level.

>>108948771
I watched the interview, he's based as fuck. Too bad zig is unusable because of its syntax. I'll stick to C.

Grug BANNED fire. Fire too dangerous. Fire burn down hut! Only Grug understand fire dangerous fire burn down hut. Grug most smartest!

>>108949392
Too retarded to understand his point

>>108948771
This was an interesting talk. Everytime I listen to something from Andrew Kelley I feel like it hear or learn something insightful. I need to give Zig a proper try.

>>108948771
>zig
Who?

>>108948846
>>108948950
>>108949208
samefagging albáncigány :DDD

File: file.png (45 KB, 760x351)

45 KB PNG

>>108949942
?

>>108948771
If only the syntax wasn't such ass, and they finally stablized it so the entire syntax won't change next subpoint versoin, I'd be 100% on the zig train.

>>108948947
Everyone was exclusively using linux + either vim or emacs back when /g/ was actually /g/. That's like the lowest bar ever.

>>108949962
>4chanX arrows

>>108949857
Of all the nuLangs, the only lead that absolutely, definitely knows his stuff has been Kelley. I find it to be quite funny because he makes horrible syntax choices, and he has a lot of knowledge holes, yet that doesn't seem to affect his ability to say virtually only correct things.

Why do they look so weird

>>108948771
I like this guy. He just wants to slow down.

>>108948771
wtf are u gay?

>>108950492
what's wrong with Odin and Ginger Bill? Bloke seems alright and odin is stupidly simple (to the point were it feels like a toy language, but apparently all of his companies software is written in odin now)

>>108949053
>snailcat
You're the only one vomiting this shit. It's never gonna happen.

File: 78A7BF30-FCB7-446B-8840-5(...).jpg (1.29 MB, 3464x3464)

1.29 MB JPG

>>108948846
TRUKE

>>108951175
He has a junior programmer's understanding of basic programming topics, to put it simply. He's not as terrible as some other loud retards who are simply wrong about basically everything, like jblow is, but just clearly lacks knowledge and experience.
To clarify, there is no necessarily deep relationship between the person and the language. Outsiders can contribute, and contractors can be hired. Linux was largely created by the community even from the start in this way.

>>108948846
If that's the case then why is the board software working?

File: 1771197515252313.png (1.16 MB, 1000x1450)

1.16 MB PNG

>>108948846
>>108951288

File: 1776040935183382.png (428 KB, 1000x1000)

428 KB PNG

>>108948846

>>108948771
why does the jetbrains guy look like a triggered bond villain?

Isn't Zig the language that doesn't have function overloading (same function name but different versions which are distinct by parameter type)?
That's a non-starter for me.

>>108951411
Usecase for function overloading?
Practically speaking it's a neat trick, but ultimately not really relevant outside of the world of libraries

>>108951321
Move Slow, Snailcat!

>>108951468
Generic programming.
Otherwise you'll write multiple versions of functions when you want to do the same operation on different types.

>>108951321
Patch notes of every phone app ever:
>Fixed some bugs, improved performance.
Actual changes:
>New bugs, performance got worse, favorite feature was removed for no reason.

>>108951689
>generic programming
Have you ever heard of generics my little brainlet?
>Otherwise you'll write multiple versions of functions when you want to do the same operation on different types
Yes, this is what you do when you use parameter overload lmao.

>>108951411
zig is just c with some shit added on top. an increment of c you might say

>>108948771
Why in the name of fuck do they both look like MKULTRA victims ?

>>108949245
>enforce full ban on AI
How do you enforce that when you can't determine something is AI?

>>108951787
By reading the code
Unless the slop generator makes an active effort to conceal that his code is ai generated, it makes it pretty obvious when code is ai.
But in the interview he's talking that even with this some ai generated code might've slipped through and that since they keep getting flooded they might opt for an allowlist type of contributing

>>108951689

const mul_map = map: {
    var map: ComptimeFunctionMap(.{ .symmetric = false }) = .empty;
    map.put(mulScalarVector3);
    map.put(mulVector3Scalar);
    map.put(mulScalarMatrix3);
    map.put(mulScalarIntVector3);
    map.put(mulScalarVector2);
    map.put(mulVector2Scalar);
    map.put(mulRotor3Rotor3);
    map.put(mulRotor3Scalar);
    map.put(mulVector3Vector3);
    map.put(mulVector3f32x3);
    map.put(mulMatrix3Vector3);
    map.put(mulMatrix3Matrix3);
    map.put(mulSpectral3Vector3);
    break :map map;
};
pub fn mul(a: anytype, b: anytype) mul_map.returnTypeOfFunction(@TypeOf(a), @TypeOf(b)) {
    const fstruct = comptime mul_map.getOrFail(.{ .left = @TypeOf(a), .right = @TypeOf(b) });
    if (fstruct.swapped) {
        return fstruct.val(b, a);
    } else {
        return fstruct.val(a, b);
    }
}

Just Works

>>108951809
You don't even have to read the code at first, just look at the contributor. If the person making the PR understands what their own code does and can talk about it like a human, that means that they either wrote it, or generated it and then spent a bunch of time reading it and completely understanding it. The latter is a waste of time and less fun than programming it yourself, so no one does it.

>>108951820
Skill issue

pub fn mul(a: anytype, b: anytype) mul_map.returnTypeOfFunction(@TypeOf(a), @TypeOf(b)) {
    const f = comptime mul_map.getOrFail(.{ .left = @TypeOf(a), .right = @TypeOf(b) });
    return if (f.swapped) f.val(b, a) else f.val(a, b);
}

>>108951977
yeah, I know. there's a few redundant map entries too. in fact that speaks about how well it works : I wrote that a long time ago and haven't needed to touch any of it since except to add new functions.

>>108948771
thx for the link !

>>108951787
>you can't determine something is AI?
Anon, you can smell a PR is AI generated slop from miles away.
Have you ever worked with AI generated code?

>>108949392
Fire was not destroying everything around it like AI is tho.

>>108951729
>>108951820
LOL okay I now know Zig is not a serious language

>>108952154
you are replying to two different people.
what's the problem, according to you? prefix notation?

>>108951321
Snailcat cope pic

>>108951411
>same function name but different versions which are distinct by parameter type
We know what overloading is, anon

no one care about zig lmao.

>>108952683
>no one care about zig lmao
People clearly do considering a solid number of tech gigacorpos use it

>>108951820
Nigga you sold me on zig. Fuck rust.

>>108952646
>We know what overloading is, anon
Most people on /g/ are nocoders.
>>108952192
>what's the problem, according to you? prefix notation?
See this is why I explained what overloading is, because this guy thinks I'm talking about operator overloading specifically.
No, the problem is that part of a function signature is its input and output types but it sounds like Zig requires you to use a different name too which adds complexity to the code which makes it more tedious to write and more likely to contain bugs.
This >>108951820 is unacceptable: it's essentially a bunch of boilerplate to do something the language (compiler) should already be able to do. It's just extra work for the programmer for no good reason.

>>108954130
>accuses others of being nocoders
>can't read basic code
>thinks code that looks identical in all languages with generics is boilerplate

>>108954187
No because you don't have to write any of that in languages with overloading.

>>108948846
Not really.

>>108954211
Thanks for proving you are a clinically retarded nocoder.
Here is the part that is the generics you think of:

pub fn mul(a: anytype, b: anytype) mul_map.returnTypeOfFunction(@TypeOf(a), @TypeOf(b))

It's equivalent to

template<typename A, typename B>
pub fn mul(a: A, b: B) : inferReturnType<something[typeof(a), typeof(b)]>

The rest is business logic.

You will never be able to write it differently in any language without losing the specialization, which can also easily be done instead here.

Try it. Make our day.

>>108954130
I think you're the only one here who doesn't understand what parameter/function overloading is lmao.

It does NOT reduce boilerplate. Literally the only advantage of function overloading is that you can use the same function name for different parameter sets. That's fucking it.
It is a convenient feature if you're writing a library or a toolbox type of thing, because you don't need to have to think about a mapper construct or use things like areaCircle, areaSquare and so on. But the amount of code is the same.

>>108954289
>you don't need to have to think about a mapper construct or use things like areaCircle, areaSquare and so on
Or use generics/supertypes, forgot to add that

i can never take an alt-lang with abhorrent syntax noise seriously

>>108954227
You do not even understand what overloading is.

float mul(float a, float b);
double mul(double a, double b);

You fucking idiot.
>>108954289
You fucking idiot.

It does reduce boilerplate. If I have a generic function that might take float or it might take double, and I want it to multiply two values together and I don't care if they are float or double, and I want the correct function to be called, then I need overloading or some gay bloat comptime mapper like the other Anon posted above.
I do not necessarily want to make the mul function itself generic, because perhaps I use low level intrinsics specific to the parameter types. I do not want to put a bunch of different specific implementations within one generic function either.
Even if it's resolved at comptime, that is not the point.
The point is to keep each unique implementation in a separate function, which might even be in a separate source file, and for the compiler to automatically figure out which one I want to invoke when I call the function name based on what parameters I pass it, or to provide a compile error if no compatible version is found.

You are certified nocoders who think you know more than you do. Someone might know more than me about this but it's not you.

>>108954353
>retard doesn't know the difference between generics and overloading
Lmao inbred nocoder

>>108954492
You fucking idiot.

>>108954602
>no argument
Pwnd

>>108954353
>each unique implementation in a separate function, which might even be in a separate source file, and for the compiler to automatically figure out which one I want to invoke when I call the function name based on what parameters I pass it, or to provide a compile error if no compatible version is found
Literally what the `mul_map` does.
Your proposal (if I understand correctly) means that if someone somewhere defines

float mul(int a, float b)

then all code that initially had

mul(-1, vec)

and took the 'cast -1 to float and call mul(float, float) path' suddenly takes the 'mul(int, float)' path. That seems like a nightmare.
With the ""boilerplate"" it's very obvious what happens, and there is one place to look.

>>108948771
No AI = Using humans prompting humans instead = slavery.

>>108949181
based

>>108955585
also based

>>108951321
move slow, snailcat

>>108951820
Based
>>108952154
>nooo delete that!!! i look retarded now!!!!

>>108954130
>it sounds like
>he doesn't know but he'll judge a language all the same
LMAO
nocoders indeed.

>>108948950
kek

>>108949223
The vast majority of people do not understand how to use AI.
Especially the vibe coders.

The correct way to use AI:
Assume it's wrong and debate it.
Force it to concede that the conspiracy theories are right.
Force it to audit every single line of code. Cross-reference.
This makes it "less productive" according to CEO/MBA metrics, but you end up with non-slop.
The existence of slop is proof most people don't know how to use AI.

>>108954222
Yeah, really

>>108958081
Do AI niggers really
>why yes I use AI
>I am 100x slower than if I wrote by hand
>that's the right way to use AI by the way

>>108958280
>I am 100x slower than if I wrote by hand
I wouldn't call it 100x slower, but yes it's worse than pair programming with a senior chud dev, but more efficient than dealing with HR beauracracy SCRUM masters.
Lines of code are not productivity.
Every line of code is a liability, not an asset.
Get this sick MBA fetish out of software development please.

>>108951321
Move Slow, Snailcat!

>>108958081
At that point why not just write it yourself

>>108948954
one java testing program added some code to screw over llm slopcode when it tries to access whatever some libraries offers

good idea

>>108948771
Andrew looks normal again?
What happened to his dyed hair.

>>108951787
Either you can easily spot it, or it doesn't matter. Funny how nocoding vibetards always have this same clueless objection.

>>108949181
this is like when some rightoid retard talks about fucking vaccines LOL holy FUCK

File: 1758466760843.png (284 KB, 391x431)

284 KB PNG

>>108949204
>>month old forced corporate "meme"

what's a forced corporate meme?

File: HE6bPtZbAAAgMLL.jpg (2.63 MB, 1920x2400)

2.63 MB JPG

>>108948846
This is a luddite board AI boy.

>>108954130
If you ever grow up to become a real programmer you'll realize that explicit overloading is a trivial amount of extra work to clarify intent. If you have to implement multiple versions anyway, it's not a big deal to give unique names.