🪅-progaming

[royal nymph]

idk but this is also crazy

[crude star]

2023

wtf

[crude star]

i kind of hate this bc it doesn't work for more than 2 columns

[crystal igloo]

Anyone know how to push changes to github on gitkraken

im having trouble uploading all my files which I was coding locally, just decided yesterday I wanted them to go on a git repo and now it wont push all the files there only the README and SECURITY

[royal nymph]

.button-grid {
  padding: 1em;
  display: grid;
  gap: 1em;
  grid-template-columns: repeat(3, 1fr);

  > :last-child:nth-child(3n + 1) {
      grid-column: 1 / -1;
  }
  > :last-child:nth-child(3n + 2) {
      grid-column: 2 / -1;
  }
}

(terrible please dont do)

becomes more complex the more columns you do lmao

[shrewd canopy]

This is why I will never use css/html

[royal nymph]

@crude star rate

<div class="button-grid-3">
    <button>Click Me</button>
    <button>Click Me</button>
    <button>Click Me</button>
</div>
<button onclick="addButton3()">Add button</button>

<div class="button-grid-4">
    <button>Click Me</button>
    <button>Click Me</button>
    <button>Click Me</button>
</div>
<button onclick="addButton4()">Add button</button>

.button-grid-3 {
  padding: 1em;
  display: grid;
  gap: 1em;
  grid-template-columns: repeat(3, 1fr);

  > :last-child:nth-child(3n + 1) {
      grid-column: 1 / -1;
  }
  > :last-child:nth-child(3n + 2) {
      grid-column: 2 / -1;
  }
}

.button-grid-4 {
  padding: 1em;
  display: grid;
  gap: 1em;
  grid-template-columns: repeat(4, 1fr);

  > :last-child:nth-child(4n + 1) {
      grid-column: 1 / -1;
  }
  > :last-child:nth-child(4n + 2) {
      grid-column: 2 / -1;
  }
  > :last-child:nth-child(4n + 3) {
      grid-column: 3 / -1;
  }
}

const buttonGrid3 = document.querySelector(".button-grid-3");

function addButton3() {
  buttonGrid3.append(buttonGrid3.firstElementChild.cloneNode(true));
}

const buttonGrid4 = document.querySelector(".button-grid-4");

function addButton4() {
  buttonGrid4.append(buttonGrid4.firstElementChild.cloneNode(true));
}

tbf you can easily generate this with scss

(sadly you can't do this with css variables)

[royal nymph]

@mixin grid-stretch($cols) {
  @for $i from 1 through ($cols - 1) {
    > :last-child:nth-child(#{$cols}n + #{$i}) {
      grid-column: #{$i} / -1;
    }
  }
}

.my-grid {
  @include grid-stretch(5);
}

[crude star]

I would expect for 3 columns the last 2 to simply be split in half
in which case you can just use flexbox for it

[royal nymph]

yeah xD

but flex for smth like this is always cursed because you have to hardcode width and account for padding, margin, gap

honestly I'm surprised there's no css property for this

it seems like a fairly common use case to want to have n columns but also stretch to fill

[crude star]

:nth-child(odd) is nice I think when you want to colour odd lines of a table darker

[crude star]

flex: calc(1/3 * 100%) and hope the margins aren't big enough to break it

it is cursed though

[royal nymph]

there's not really clean css solution

with a bit of js you can do a clean solution

combining grid + flex

[deep mulch]

gooddd

rosie will still somehow win

[hoary sluice]

@valid jetty

why does elle not compile

yuore using reserved keyword gen

is there a cargo option to bypass that

[lucid trail]

Edition 2024?

[hoary sluice]

i think its supposed to be 2021

but i get this

even tho

its allowed

im using cargo 1.92

[lucid trail]

nix fixes this

[hoary sluice]

what

wdym

[deep mulch]

@hoary sluice 🦅ly

[hoary sluice]

@deep mulch er

@valid jetty @lucid trail how do i run elle

do i need cargo <1.88

do u have a devshell for it

[lucid trail]

idk i'm just using

cargo --version
cargo 1.88.0-nightly (d811228b1 2025-04-15)

[deep mulch]

secret

opening a pr to elle with 5000 deletions 20 additions

[hoary sluice]

and how did yuo get 1.88

[lucid trail]

i think i used rustup

[hoary sluice]

rustup 🤢

[lucid trail]

yeah i know

[hoary sluice]

ill pr a dev shell

hope rosie accepts it

[valid jetty]

elle has a rust version lock

https://github.com/acquitelol/elle/blob/rewrite/rust-toolchain.toml

[pearl stag]

rust-toolchain.toml:

[toolchain]
channel = "nightly-2025-04-20"

[valid jetty]

your nix is stinky

install this toolchain

[errant granite]

Hello can i link my instagram account to discord somehow?

[valid jetty]

rust is fucking stupid

on 1.90.0 they stabilized let chains

but only on rust 2024, which i can't use because that also RESERVES the gen keyword

they don't use the keyword, they just reserve it

[hoary sluice]

yes

i noticed

[pseudo sierra]

insane

[valid jetty]

so i have to stay on rust 2021, which no longer has let chains on rust newer than 1.88.0

it's fucking deranged

[hoary sluice]

why cant you just replace usages of gen

i had this problem too

[valid jetty]

because there are many and i don't want to do that refactor right now

[hoary sluice]

when using get instead of generation

[valid jetty]

i also don't want to rename it, gen is a good name

[hoary sluice]

do u not have a refactor macro

in zed

[valid jetty]

each codegen component is given 3 things, self (the astnode), ctx (stuff like the current function, module, cascading type, whether it's within a a return statement) and gen (stuff like tmp counter, scopes, address map, etc.)

gen stands for codegen, but codegen is too long

gen is the perfect name for that variable

[hoary sluice]

@valid jetty

are u home rn

[valid jetty]

yeah

[hoary sluice]

do cargo clean && time cargo build --release

on elle

[valid jetty]

lmao okay

[hoary sluice]

wait the time is unneccessary

rust tells u

[lucid trail]

there's a command to migrate it for you i think

[valid jetty]

there is

but i dont want to think of an alternative name

[hoary sluice]

generator?

[lucid trail]

r#gen but its so ugly

[valid jetty]

its not a generator

[valid jetty]

lmao

crazy

what the fuck just happened

[lucid trail]

so slow

[hoary sluice]

whats yalls cpu

[lucid trail]

m3

[valid jetty]

m1

[hoary sluice]

mac 🤮

i dont think running elle in 13 seconds was wroth the 700$

[lucid trail]

uhg i cant buy a new mac until liquid glass is over

cause they all come with macos 26

[deep mulch]

ryzen 7 9700x

[hoary sluice]

try compiling elle

[deep mulch]

okay

acquite LOL!!!

works

[hoary sluice]

no like

do

cargo clean && cargo build --release

[hoary sluice]

nice

i wonder if m5 is faster than that

[lucid trail]

they should do more rust compilation benchmarks

[deep mulch]

17.96s now

[hoary sluice]

what did u change

i wonder if i can get it under 10s

if i put my fans to max and overclock to like 7ghz

meh thats probably not possible

its rated for 5.76

[deep mulch]

just reran

does cargo build use all threads

[hoary sluice]

idk

yes

[deep mulch]

how

[hoary sluice]

RUSTFLAGS="-Z threads=32" cargo build --release

[deep mulch]

love

elle build speedrun

[hoary sluice]

how????

[deep mulch]

idk

[valid jetty]

@lucid trail i want to add a range of builtin methods to static arrays so that you can use them like vectors and matrices but i dont have function overloading

[deep mulch]

@valid jetty how fast does elle clean build for you

[valid jetty]

i was thinking to call the vector version of the function vname and then just name for the matrix ones

T[A][B]::mul() and T[A]::vmul()

[hoary sluice]

@valid jetty https://github.com/acquitelol/elle/pull/8

[hoary sluice]

what command did u use

why does it run faster than for me

[valid jetty]

i thought you switched back to kde

[deep mulch]

this but 16 threads

[hoary sluice]

this has nothing to do with kde

u can have kde on nix

kde is a desktop environment

nixos is an operating system

[valid jetty]

oh wait no i meant i thought you switched back to a sane distro

[deep mulch]

i also set power plan to performance

[hoary sluice]

also no im on hyperland

[hoary sluice]

nixos is a sane distro

[deep mulch]

@valid jetty fixx

its not an error

[hoary sluice]

ok im overclocking my cpu just to compile elle

[valid jetty]

mv wrote that into stderr

so make thinks its an error

[deep mulch]

i think it just needs some flag

[valid jetty]

true

[shrewd canopy]

No its unsane

[valid jetty]

is this cursed

fn __arr__::__iter__<T, N>(T[N] self) {
    return Array::slice_from_raw_parts(self.len(), #cast(T *, self)).iter();
}

external fn __arr__::__iter__<T, N>(T[N] self) @alias(__arr__::iter) -> Iterator<T, DoubleEnded>;

[formal belfry]

long enough id skip over when reading it

[lucid trail]

ah i really see the benefit of function overloading now

i'm imagining it can't be that hard right

[valid jetty]

even with function overloading, it would be an ambiguous case without generic constraints

because T1[A1][B1] fits into T2[A2]

its just T2 = T1[A1] and A2 = B1

so if you had function overloading for T[A] and T[A][B] you wouldnt be able to guess which function to call

[lucid trail]

oh you’re right

[deep mulch]

very

[valid jetty]

zoot doesnt know what pointers are

[lucid trail]

yeah you’d want traits here

hmm

[valid jetty]

actually would this code even be valid

im concerned because.. it works

maybe it should accept the static array by pointer

actually no i see why that works

it only blits the data when the static array is the return value of a function

in the __iter__ method its essentially moving the static array into the iter function from the caller

there are no copies being made

ummm ok idk how i feel about that behavior

i may make it copy when a static array is accepted by value into a function too

ok ill think about this tomorrow im way too tired

[deep mulch]

elle syntax so evil

i will rewrite

and pr

@valid jetty willmerge

[hoary sluice]

@deep mulch i was running on the vcache ccd

which is much slower

[deep mulch]

o

how much faster is it without

[hoary sluice]

checking rn

@valid jetty @deep mulch

this is running on all 16 cores

half of them are 5.76ghz

other half are 4.2ghz

this is without overclocking

im too lazy to set up overclocking now

already wasted an hour on this

cpu peaks at 74 degrees

@valid jetty merge the pr

[dense sand]

Anyone used a custom vitest runner here

[shrewd canopy]

Does elle even support windows 🐟

[hoary sluice]

hopefully not

[shrewd canopy]

Hopefully it does

[lucid trail]

Oh I see, we usually think of T[4] the same as T[4][1]

you could coerce them into each other

[fleet cedar]

Who is this "we" and how can they possibly be this wrong?

[lucid trail]

i meant humans when thinking of 1d matrices and vectors

[fleet cedar]

Mathematicians, and linear algebraists in particular, are indeed typically very bad at types

[lucid trail]

yeah

well a vector is an element of a vector space

so its different

[frosty obsidian]

flagging added

[frosty obsidian]

stylistic changes

[frosty obsidian]

it goes from blue to green to yellow/gold to darker and darker reds

8 is a very dark red

basically just gets warmer

[deep mulch]

nop

someone already made a compose mine sweeper it looks very similar @frosty obsidian

[frosty obsidian]

yeah i know

[valid jetty]

i'll merge after i push all my static array type work

there are like 8 commits incoming

[fleet cedar]

Do hexagons or triangles instead of squares to make it a little more unique

[ornate quiver]

make the seed nicer

actually hide it until the game is over

[frosty obsidian]

that's the plan

its only there for debug reasons

[deep mulch]

wing forgot

[dapper notch]

Yn

[nimble bone]

Cursor @frosty obsidian

[fleet cedar]

Ew, non-monospace ligature

[winged mantle]

intellij vim plugin is pretty decent

[tired vigil]

I think it is monospace, just compressed into 1ch width
which means that the cursor movements will be fucked up

[fleet cedar]

If the string fi is not the same width as fo, it's not monospace

[tired vigil]

hmm true ig

[fleet cedar]

fi should naturally be one cell wide of course

[lucid trail]

oh my god parsing * vs ** vs *T vs **T is going to be horrible isn't it

[valid jetty]

yeahhhh

it's only not ambiguous if all cases in the syntax where a type may be are different from the cases where an expression can be

[lucid trail]

i think i'm gonna use ~T

[lucid trail]

never vibecoding with Opus 4.1 thinking again

[supple whale]

did you mean?

[undone wolf]

hi, im looking for python developers.
paying $10k for py job.
dm me for pof and more details pls

[lucid trail]

ok yeah i can parse this now, it's not ambigious because it's clear where types are allowed and where it isnt

[valid jetty]

@young flicker

[jade stone]

Nope

It has support for like zero builtin vim commands

So making a custom config is borderline impossible if you're doing anything but remapping keys

[stoic mauve]

https://projecteuler.net/problem=949 drops this and leaves

[supple whale]

OH GOD NO

MATH

ALSO LEAVES

[deep mulch]

@valid jetty hii

[nimble bone]

hi, im looking for python developers.
paying $10k for py job.
dm me for pof and more details pls

[nimble bone]

DM @young flicker

or @jade stone

[jade stone]

husk

python sucks

[shrewd canopy]

nop

python good

[nimble bone]

or @frosty obsidian

[jade stone]

sorry i hate being forced to indent my code with four spaces

tabs >>>>>>>>>

[shrewd canopy]

you can use 1 space too

just use it consistently

[ionic lake]

why do you guys keep complaining about indenting

just write code normally

[drifting yarrow]

yesterday on github i found something interesting..
https://github.com/oliver-ni
^ from this profile when i click on their "pokétwo repo" it takes me to ..poketwo/poketwo and not oliver-ni/poketwo

is it a github organization? why is that happening?

[jade stone]

you can pin org repos

[drifting yarrow]

ahh, so its an different org repo pinned to the "oliver-ni" acc?

[jade stone]

yes

[drifting yarrow]

i was thinking some binding shi 💀
made under oliver-ni then redirected to poketwo
( dont even makes sense idk what i was rthinking lol )

[drifting yarrow]

tyyyy

[jade stone]

i think that could happen if the repo was moved to the org

[drifting yarrow]

ooh, interesting

[ionic lake]

its the same without indents

not without indents i meant braces code like js

[winged mantle]

wait does python force spaces

awful

tabs work on my machine

[valid jetty]

https://x.com/FFmpeg/status/1986716471297495392

FFmpeg (@FFmpeg)

The most interesting and accurate technical responses on X come from anonymous accounts with cartoon profile pictures, pictures of toast or other generic items.

The one with corporate headshots and ex-FAANG in their bios usually provide nothing of value.

X · Nov 7, 2025, 8:44 AM

[valid jetty]

it enforces one or the other, not both

try to mix tabs and spaces

[crude star]

https://x.com/FFmpeg/status/1986720239887757416

FFmpeg (@FFmpeg)

@Aut4rk Why are you converting FFmpeg's nice Intel ASM syntax to ugly AT&T syntax?

X · Nov 7, 2025, 8:59 AM

[shrewd canopy]

Based

Intel syntax > AT&T

[deep mulch]

true

[stark timber]

hi

[jade stone]

esbuild so bad

[shrewd canopy]

?

[jade stone]

should be minified to function*foo(){yield}

?remind 4hr learn go and pr minification to esbuild

[delicate grove]

Alright @jade stone, in 4 hours: learn go and pr minification to esbuild

[jade stone]

both terser and swc do this and from what i can gather from reading mdn it's correct and unobservable

[shrewd canopy]

👉 https://github.com/evanw/esbuild/issues/create

[jade stone]

@deep mulch typescript AST will make me go insane

[supple whale]

XD

[supple whale]

afaik minify isnt esbuilt but terser

and afaik terser is context-aware

instead uses ast and does keyword replaces

[jade stone]

that makes no sense, what?

[supple whale]

so it assumes there might be shit after the yield

and forcefully yields something

[supple whale]

esbuuild runs terser to minify shit

no?

[jade stone]

no??????????????

[supple whale]

yes

[jade stone]

you are 100% wrong

esbuild is 100% written in go

terser is written in js

also they are different programs

esbuild is a bundler

terser is a standalone minifier

[supple whale]

ah they use terser

but only for testing against it

[jade stone]

terser is a bit better than esbuild, however a good amount slower

[supple whale]

apple accidentally sourcemapped all their code on apps.appstore a while back, and are now DMCA'ing everything that republishes it
https://github.com/minhducdz99/apps.apple.com
if some1 wants to grab it to learn here it is

[delicate grove]

@jade stone, <t:1762528130:R>: learn go and pr minification to esbuild

[winged mantle]

😇 I would never do such a thing

[valid jetty]

@hoary sluice @deep mulch my github about me is finally real elle code

well the english part was real elle code too but now the jp part can be entirely jp

[deep mulch]

yappanese

[valid jetty]

instead of just half of it

[lucid trail]

was this not previously not possible

[valid jetty]

it was possible for a while

its been a thing for 4 months

but i just forgot about it in my bio

[deep mulch]

is that crylic

russian

[valid jetty]

yeah

english, japanese, russian, australian, and emoji

[deep mulch]

@valid jetty make roiescript

[ornate quiver]

yapinglish @deep mulch

[paper scroll]

do yall think a pr would be merged if i moved vesktop tray to libvesktop instead of electron on linux

dont really care to make a pr if there is never a chance itll be added

[pseudo sierra]

maybe?

[paper scroll]

[shrewd canopy]

No

No point of doing so

[paper scroll]

i mean on x11 plasma and some others theres a huge memory leak for electron tray icon setting

just electron bug but

[shrewd canopy]

Go make a issue on Electron Github

[paper scroll]

[jade stone]

Ask in #🖥-vesktop-development

[paper scroll]

weird had that channel hidden, also i cant speak there lol

[hoary sluice]

RUSSIAN ELLE

[jade stone]

Make a mod mail and ask for programming role

[hoary sluice]

how is this server not termed

i thought mods were against tos

@deep mulch give me talking in vencord dev channels prems

[royal nymph]

haram

[ionic lake]

yes

[vagrant crescent]

harami

[fleet cedar]

Salami

[winged mantle]

https://fixupx.com/tsoding/status/1987214358838190137

Тsфdiиg (@tsoding)

I removed ';' from C

**💬 11 🔁 11 ❤️ 106 👁️ 2.4K **

▶ Play video

FixupX · Nov 8, 2025, 5:43 PM

[valid jetty]

already seen

[winged mantle]

obvs you're in the tsoding discord

[valid jetty]

oh youre in there too

crazy

[deep mulch]

@valid jetty hiii

[valid jetty]

they optimized this game so much since i last played it

the graphics are beautiful

@young flicker

[hoary sluice]

do people use warp unironically

or wave

[valid jetty]

i tried it

[hoary sluice]

its not even a terminal atp

i tried wave

[valid jetty]

ever since they added ai i got rid of warp so fucking fast

now i use ghostty

[hoary sluice]

when yuo enter vim on wave its not fullscreen theres an ai sidebar

its like

on all sides its not fullscreen

[royal nymph]

ghostty is horror

[hoary sluice]

i tried ghostty but didnt like it

kitty is better for now

[royal nymph]

I tried ghostty but something was very bad about it I forgot what so I stopped using it

also no packages for debian and rpm based distros is unreal

[hoary sluice]

@valid jetty whats the appeal with ghostty

i dont like it when apps have gtk title bars and configs unneccessarily

[royal nymph]

made by funny twitter guy

[valid jetty]

nothing in particular i just have a minimal config and ive been using it for a while without problems

[hoary sluice]

look guys

i need a faster terminal

kitty is too unoptimized

[shrewd canopy]

xterm

and u must use twm and X11

or 🪟 and cmd.exe :)

[royal nymph]

#🌺-regulars message

#🌺-regulars message

basic features don't work in ghostty

[hoary sluice]

yea i remember that

there was something

not working

[shrewd canopy]

what features

[hoary sluice]

look at the link

s

also the title bar is annoying

[shrewd canopy]

I cant read regulars channel

[hoary sluice]

yes you can disable it

but it shouldnt be there in the first place

ghostty is not made for people on a twm

[hoary sluice]

why x11

[hoary sluice]

why xterm

[shrewd canopy]

stable for ages and always works fine

[shrewd canopy]

cuz its fast

[hoary sluice]

ok

4.4bsd is also stable for ages

and works fine

are you gonna daily drive it now

[shrewd canopy]

sure if all of games/software/drivers worked on it

[hoary sluice]

all games and software from x11 work on wayland

the only thing that doesnt work is x11 tooling like xclip but theres wayland tooling

[shrewd canopy]

Xfce doesn't, at least for me

[hoary sluice]

ok thats x11 specific software

bspwm doesnt work eihter

neither does hyprwm

or i3

or banana

[jade stone]

How

Never been slow for me

[hoary sluice]

its a joke

look at the video

[jade stone]

Yeah I saw it but I didn't know that was Kitty in the video

[hoary sluice]

i thought it was obvious

[jade stone]

Ok

[solid gazelle]

Heyy

is there a way to disable these in Astro

or make them shorter 😅

[nimble bone]

no

they allow you to use per component styles

[royal nymph]

why

theyre for scoped styles

[solid gazelle]

very long

[royal nymph]

you can change it to use classes https://docs.astro.build/en/reference/configuration-reference/#scopedstylestrategy

but it's no difference

[frosty obsidian]

why care

you're not supposed to look at the final html

[dawn ledge]

youre embedding styles into your elements surely you dont care about a few extra bytes

/shrug

where'd my reply go

[dawn ledge]

.

although if youre using something like river wm that doesnt support the ssd protocol gtk uses (FUCK GTK) youre stuck with ugly borders

[deep mulch]

@frosty obsidian

[valid jetty]

im not typing all of that again

[deep mulch]

bloattt

elle so bloated

remove

@valid jetty Elle needs rewrite

[valid jetty]

its not that bloated

there are maybe 7000 lines of code max that i could genuinely get rid of if i refactor the parser and compiler

the rest is kinda necessary

[deep mulch]

wrong

I rewrite Elle in 200 lines

make Elle faster @valid jetty

[lucid trail]

oh some elle benchmarks would be interesting

[strange matrix]

(ping becos you have rust icon)

can you explain

while let Some(x) = x {
}

?

i've tried reading the book
i really can't understand it

[hoary sluice]

let mut iter = vec![1, 2, 3].into_iter()

while let Some(x) = iter.next() {
    do_something(x);
}

do you understand it like this?

its the same as

for x in vec![1, 2, 3] {
    do_something(x);
}

[strange matrix]

i don't get the part Some(x) = iter.next()
is this a check? what is this

[hoary sluice]

do you know what if let does

[strange matrix]

also no

[hoary sluice]

read that first

[strange matrix]

okay hold on

okay i don't understand this either 😭

i just understand that i can use it to check enum values

[strange matrix]

is this PDFLinux or something else ?

oh its WASM

[fleet cedar]

if let pat = expr { ... }
``` is shorthand for ```rs
match expr {
    pat => { ... }
    _ => {}
}
``` (`else` version should be obvious)


```rs
while let pat = expr { ... }
``` is shorthand for ```rs
loop {
    match expr {
        pat => { ... }
        _ => break
    }
}```

[strange matrix]

guessing else is the _ => {}

thank uuuu

do you want to check my project?

[fleet cedar]

No

[strange matrix]

okay

[hoary sluice]

have u learned how enums work

and what result/option is

[strange matrix]

i think i know 😭

[hoary sluice]

i mean like the rust specifics of enums

tuple values, struct values, etc

[strange matrix]

probably not

i've probably learned them before
just haven't used

so i forgor now

[hoary sluice]

learn that before learning while let

[strange matrix]

'k

[lucid trail]

what the hell

[fleet cedar]

Oh, learning about provenance for the first time?

[valid jetty]

well consider this

int xs[5];
int *foo = xs + 5;

int ys[5];
int *bar = &ys[0];
``` if they just happen to be contiguous in memory and `xs + 5` is the same memory address as `&ys[0]`, are they the same pointer? no!

[solid gazelle]

explain

[valid jetty]

Array#includes will iterate through every element in the array, checking items[i] == item

[lucid trail]

yeah

[valid jetty]

Set#has will calculate the hash of item and then check if there is something living at that place in the array

idk what js's set is using

most likely its using a static array of buckets which are linked lists

[solid gazelle]

waa

[lucid trail]

yeah this is exactly the example

an quite interesting problem

[valid jetty]

basically the way that a hashmap/hashset works is:

• calculate a ""hash"" of the item
• mod by length of your array
• put the item in that place in the array
• if there are multiple things that hash to the same thing, put them in a linked list at that index

well, at least for open addressing

and then essentially, you assume a hypothetical scenario where each item perfectly hashes to a unique slot in the array and there are 0 collisions

therefore that makes the operation technically O(1) since calculating the hash of the item is not dependent of how many items there are in the set

unlike an array

i hope that makes sense

[solid gazelle]

thank u

[valid jetty]

im not sure what js uses for its maps and sets

but open addressing is kinda the most common nowadays i think

closed addressing is basically where you have a dynamic array instead of a static array of linked lists

and then you calculate the hash, and if there is a collision, you go to the next slot in the array until there is an empty one

if you run out then you grow the array

this is called linear probing

but the problem with this approach is that if many things hash to the same thing you will end up with loads of collisions and so fragmentation

and then that fragmentation will lead to slow inserts and slow retrieval, eventually approaching O(n)

im not sure exactly how quadratic probing is designed to work but im sure theres a similar problem

[hoary sluice]

you can switch them after this code

so they are the same pointer

[valid jetty]

because imagine there is an array like

[ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ]

and you wanna insert 5

5 hashes to 13

so you insert it

[ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][ ][5]

now you wanna insert more things...

and you end up here

[ ][12][22][7][15][16][-1][92][26][54][19][3][14][5]

and you wanna insert 192

that hashes to 1

but there is already an item at that slot, so you go to the next one .. and repeat that until you end up at the very first slow which is empty

..essentially O(n)

[valid jetty]

its undefined behavior

[valid jetty]

..in case it wasnt obvious i implemented my own closed addressing and open addressing hashmaps and hashsets in elle lmao

the garbage collector uses an open addressing hashmap, the actual hashmap in the stdlib uses a closed addressing system

there are loads of good videos on youtube

yeah this is an open addressing hashmap

a static array of linked lists

[solid gazelle]

rn like react native

[valid jetty]

you mean "never gotten into"?

@hoary sluice @deep mulch https://www.youtube.com/watch?v=htYh-Tq7ZBI

deriving geometric algebra using vector multiplication!!!!!!!!!

links complex numbers and quaternions using bivectors

fucking awesome lecture

im in cs rn

[lucid trail]

i love her channel

so good

[solid gazelle]

ahhhhhhh i udnerstand now

rosie is THE smartest

[valid jetty]

yeah but my grades are bad

[lucid trail]

oh which unis do you want to go to?

[valid jetty]

because i was uhhhhhh......... not in the best mental state during my mocks last year

i have AAC right now

[valid jetty]

edinburgh or leeds

[lucid trail]

edinburgh is a nice place, i visited

they also rejected me

[valid jetty]

yeah i heard you wont even get an offer unless you have 3A* predicted

which is why im hoping my personal statement and personal circumstances in my reference are enough to get a contextual offer

since its hopefully clear im passionate about cs

well i am applying

[solid gazelle]

we don't do the thing of going to good unis where i live

[lucid trail]

I feel like AAA is not that hard, but thats insane

[valid jetty]

tnixc what do u think of my personal statement

well thats not all of it hang on

[lucid trail]

mine was so shit 😭

I don't think says a lot about yourself as a person, more of what you did. like ok you did cool things but why like compilers? or why did you find deriving geometric algebra from first principles so interesting?

[valid jetty]

yeah im gonna put my extracurriculars into my reference rather than my personal statement

i love playing rhythm games and listening to music but i ran out of characters

[lucid trail]

i think in particular focus on what excites you about the subject

[valid jetty]

i mean i did in the first paragraph

well theyre forcing you to structure it really horribly

[lucid trail]

also keep in mind the admissions people are not necessarily technical people, theres quite a bit of jargon

[valid jetty]

i have to answer these in order

Why do you want to study this course or subject?

How have your qualifications and studies helped you to prepare for this course or subject?

What else have you done to prepare outside of education, and why are these experiences useful?

[valid jetty]

idk i was told it is best to include as much technicality as possible in it

the more complicated things there are the more obvious it is that youll succeed at the course

to quote my form tutor quite literally "i have no idea what any of these things mean, and thats very good"

[lucid trail]

Interesting, I have no evidence to say otherwise

[valid jetty]

and thats also the sentiment that various other people ive asked have told me

and anyway its not really technical jargon

someone who does computer science at university should know what most things there are

except maybe for like, hermes bundles and monomorphization?

but the other stuff is all pretty standard stuff

[lucid trail]

thats true

some of my classmates have not heard of UTF-8

[valid jetty]

like there is no way an admissions person doesnt know what a garbage collector or lazy iterator is

or a flood fill algorithm

[lucid trail]

why would you assume they are studying cs

[valid jetty]

well im assuming since im applying for cs i would be reviewed by an admissions officer who is specifically fluent in cs

otherwise to satisfy the admissions officers you would have to make your personal statement as bland as possible and that is not what ive been told by anybody

thats why its not long blocks of text

idk i thought i structured it pretty well

at least compared to initial drafts

which were like 3 paragraphs all with like 2 sentences per paragraph

i loooooved to have super long sentences seperated by a comma

this is a uk uni

everyone in the uk needs to follow this format

[valid jetty]

you mean this?

yeah, thats true

i dont know

i dont like it either

its achronological

since i made ichigo after elle but i need to talk about it before elle

[final garden]

Achronological means out of chronological order; not arranged in the order events actually occurred.

Gemini

[valid jetty]

lmao

actually

can you ask your teachers on monday how many technical terms they understand out of that personal statement

im really curious

husk

[lucid trail]

i think you're following this too strictly

after all, I didn't, and i got in 4/5

[valid jetty]

i dont like that youre including your personal website as one of the projects on your personal website

feels like youre fluffing up an otherwise lack of projects

also why do the buttons take 5 years to take you anywhere 😭

the little circle is when i click btw

genuinely i thought my internet died when i clicked on the about button and it did nothing

the first time i clicked it, it genuinely took 10 seconds to take me to the page

[lucid trail]

it takes 2.5 seconds

[valid jetty]

seems to be getting faster each time i switch tab?

very strange behavior

WHY DO YOU NEED SOLID FOR A PERSONAL WEBSITE 😭

it should be static html

[lucid trail]

you can have static html with components too 💔💔

nothing to do with those

[crude star]

you hate interactivity

[valid jetty]

the fact that it is lagging on your phone should be telling you something

[crude star]

its called a personal website because rosie personally chooses what you should be doing

[valid jetty]

i do

at least for a personal website

for a website where you can look at some text and click on some buttons you dont need a whole ass framework

because thats essentially what a personal website is

a description of you and the things youve done, and buttons to learn more on other websites

[crude star]

?

a personal website is whatever you want it to be can u hear ur own words

[valid jetty]

whatever..

[crude star]

https://neal.fun/ i am a developer from India and this is my personal website

should i replace with static html ?

[valid jetty]

thats what im saying :3

i should make an entirely interactive website without any js

not even without any framework, no js at all

you can probably do it with radio buttons

[crude star]

css audio when

[valid jetty]

@hoary sluice lmao

[crude star]

millions will stream html

[hoary sluice]

is this elle

icy

[valid jetty]

yes

[deep mulch]

you understood everything in this

[valid jetty]

i understood most things, yes

its not really that in depth..?

[royal nymph]

help

i dont mean to flex but I got the vegetable one right first try

[hoary sluice]

arent those synonyms

[lucid trail]

cs is so annoying, so tedious

yeah i have to take cs courses as a math major

i dont think its gonna lead to getting a job either way

[hoary sluice]

both are the same on the resume

[deep mulch]

nerd

[valid jetty]

i thought that was obvious by now

i literally speak in the tsoding server

[deep mulch]

idk what that is

[shrewd canopy]

Slop environment

[valid jetty]

you know the guy on twitter who likes dynamic arrays and programming in C

he has a discord server full of nerds

[deep mulch]

no

i dont interact with twitter

@valid jetty https://www.youtube.com/watch?v=IjIFCyeo47g

i should join then call rosie a nerd and leave

[fleet cedar]

Such a silly question, of course you should

[valid jetty]

you wouldn't be able to

you need to subscribe on twitch or get vip to talk there

[deep mulch]

[valid jetty]

this guy

because he's russian..? are you just racist or..

[solid gazelle]

sorry

nvm

[solid gazelle]

im not racist err meant to ask wheres his accent from yk sorry

[ocean hatch]

libffi the hidden gem

[winged mantle]

i love how he uses makefiles for rust

i wanna do that with c# and java

[valid jetty]

@deep mulch i wrote a teeny ascii art shader

its very teeny

[runic sundial]

Make your programming language transpile into GLSL

devilish

[worldly sigil]

[crude star]

spir-v target soon

[lucid trail]

this is probably not that hard surely, translate it to SPIR-V and you could do metal too

[lofty iris]

has anyone made a midi file parser

[runic sundial]

Spirv won't work on macos

[runic sundial]

this is just painful

[hoary sluice]

being racist against russians is ok tho

[hoary sluice]

@valid jetty all of university cs

[shrewd canopy]

Tsoding isnt russian

[hoary sluice]

true but the only time ive seen microservices poorly implemented is oop

[hoary sluice]

yes he is

[shrewd canopy]

I doubt it

He does speak english but can write in russian

[hoary sluice]

here is a 2 hour stream fully in russian https://youtube.com/watch?v=f4EOMbmIqqw

and he doesnt have an accent

only a bit of rhotacism or an accent where the r is pronounced differently

[shrewd canopy]

Hm okay

I'll watch after game

[hoary sluice]

do u speak russian

oh ur ukrainian

[hoary sluice]

this was sarcasm btw

[winged mantle]

he said in a recent stream he never left russia

[valid jetty]

yeah that's why he doesn't get any money from donations and can't access most webpages

kinda sucks

[hoary sluice]

theres no way tsoding is unable to set up a donation link

[valid jetty]

nono he is able to

but i meant twitch donations from subs and patreon don't reach him

they are purely ways for him to moderate his discord server and chat lol

[winged mantle]

i should pay to get access to his server

my cursed ideas would probably be right at home

though it's not like in other cases where even if i don't talk much i know i'm supporting one of my favourite channels :(

[jade stone]

@nimble bone working on a project where i need esbuild, rollup, and vite at once

[nimble bone]

how

[jade stone]

i think ill finish tonight ()

[deep mulch]

salad addicted to web dev

[pseudo sierra]

salad insane

[jade stone]

This isn't web dev

[solid gazelle]

am i crazy or doesnt vite use esbuuild and reollup ourgh

[jade stone]

Yes but I don't mean it like that

[solid gazelle]

STINKY SADAN

[nimble bone]

https://tenor.com/view/deepseekvschatgpt-gif-16222292642975841075

@jade stone Satan

[solid gazelle]

https://cdn.discordapp.com/attachments/1181640097888747633/1437579798594457871/attachment.gif

[deep mulch]

horror

[solid gazelle]

@valid jetty inko language added web socket

[jade stone]

i made a terrible vite/rollup plugin

i kept having so many things that i just wanted to do at build time instead of runtime, so i made a plugin that would just make it easy to generate modules

and it also generates a d.ts for you

[deep mulch]

export declare const

public static void main

typescript is java

[jade stone]

export default class extends

[ornate quiver]

https://vxtwitter.com/lauriewired/status/1987994801485357139/

LaurieWired (@lauriewired)

Shader systems are ridiculously powerful if you’re clever enough.

Most people use them to create visual effects. You know what’s cooler?

Running Linux.

Inside an emulated RISC-V CPU. Inside a pixel shader. Inside of VRChat...

vxTwitter / fixvx · Nov 10, 2025, 9:24 PM

[hoary sluice]

im russian

[shrewd canopy]

And how slow is it

[pseudo sierra]

About ~250 kilohertz on a 2080Ti.

[shrewd canopy]

💀

So not even XP would run on that thing

[keen sedge]

how

[hoary sluice]

i was joking

[keen sedge]

o

ight

[autumn sedge]

w google
-# https://android-developers.googleblog.com/2025/11/raising-bar-on-battery-performance.html

[hoary sluice]

@valid jetty why have i been using almost exclusively rust outside work for 2 years and have still never had a use for lifetimes

ive used lifetimes but always ended up refactoring them out

[winged mantle]

software needs to be more complex

make things needlessly complex

[hoary sluice]

@valid jetty is this correct ```rs
pub fn map<'a, U>(&'a self, f: impl Fn(&T) -> U + 'a) -> impl Iterator<Item = U> + 'a {
self.data.iter().filter_map(move |opt| opt.as_ref().map(&f))
}

    pub fn indexed_map<'a, U>(
        &'a self,
        f: impl Fn(usize, usize, &T) -> U + 'a,
    ) -> impl Iterator<Item = U> + 'a {
        self.data
            .indexed_iter()
            .filter_map(move |((r, c), opt)| opt.as_ref().map(|v| f(r, c, v)))
    }

this is so unreadable lol

ok the lifetime is yet again not needed

@valid jetty @fleet cedar what do i call an in place map

on a grid

[pseudo sierra]

unless you're doing async/mt you almost never need lifetimes and even then 'static fixes it 99% of time

[hoary sluice]

ive made enough async stuff

and still have never used lifetimes

[fleet cedar]

map_inplace?

[royal nymph]

In Ruby you'd name it map!()

[hoary sluice]

@valid jetty u need to participate

https://codingcontest.org

at least online

[valid jetty]

@hoary sluice I LOVE DUCK TYPING

where is zoot

[valid jetty]

this is a really busy couple of weeks for me

[shrewd canopy]

left cuz #🖥-vesktop-development message

[lucid trail]

is it just that you assume T is iter

[valid jetty]

yea

well no, T is not iter

i assume T implements a method that can turn it into an iter

[fleet cedar]

I love c++ post-mono errors

[jade stone]

@valid jetty does Elle have inheritance

[pseudo sierra]

does elle have sadan exploder

[fleet cedar]

Does elle have ub

[valid jetty]

no, thank god

[jade stone]

Add

[valid jetty]

there is polymorphism if you're a bit creative

use std/prelude;

struct Square { i32 length }
const Square::draw = fn(Square *self) (0..self.length).for_each_with(fn(_, self) $println("*".repeat(self.length)), self);

struct Triangle { i32 length }
const Triangle::draw = fn(Triangle *self) (1..=self.length).for_each(fn(i) $println("*".repeat(i)));

struct Shape @nofmt {
    void *base,
    fn(void *) draw
}

fn Shape::from<T>(T *shape) {
    return Shape {
        base = shape,
        draw = T::draw
    };
}

fn Shape::draw(Shape self) {
    cb := self.draw;
    cb(self.base);
}

fn main() {
    square := Square { length = 5 };
    triangle := Triangle { length = 4 };
    shapes := [Shape::from(&square), Shape::from(&triangle)];

    for shape in shapes {
        shape.draw();
        $println();
    }
}

[jade stone]

Insane

[ionic lake]

show your code

do

[dense sand]

guys any tips for lightweight, portable rust frameworks for backend development? i need just a simple thing to expose some routes, important is thats its wasm compilable

[ionic lake]

axum? here's an example: https://github.com/tokio-rs/axum/blob/main/examples/simple-router-wasm/src/main.rs

[pseudo sierra]

axum is nice

[dawn ledge]

its not lightweight but its definitely nice

i use it

[ionic lake]

just don't pull everything from tokio and youll be fine

[dense sand]

do you guys think it would be possible to run a simple api from rust using wasm?

and call it like from typescript in node/bun

[pseudo sierra]

yes

I actually saw a repo just for that a few days ago when I was looking for something else

[dense sand]

do you think axum would work for this

[fleet cedar]

Big fan of saying api when you mean web server

[ionic lake]

are you asking for an web server api, or you want to expose something from rust to js?

for the latter, napi-rs works great, you just annotate your functions with #[napi] and it does the rest with napi_build

bun and deno have ffi so that's another sleeker option, but more annoying to work with and no node support

[dense sand]

Well it is an api - its a web server which i want to serve/manipulate data, therefore an api

[fleet cedar]

Yes but "an api" means literally nothing

[ionic lake]

a pee eye

[jade stone]

guhhhhhhhhhhhh

thanks windows

how

patching my own npm package because i forgot to push it to github 😭

how

i hate windows

GOD FORBID V8 FIX THIS BUG

at least its throwing an assertion instead of segfaulting

[supple whale]

skill issue

[jade stone]

i opened an issue with node for them to backport the fix

still not merged

[pseudo sierra]

use newer node then

[lucid trail]

I really want anonymous unions in rust, is there a way for me to do something like that

don't really want to create a type just for one function

[lucid trail]

this syntax is cool, I haven't heard of @ in rust before

[pseudo sierra]

[supple whale]

node is very behind in v8 versions

compared to chromium

[tired vigil]

Whaa what does that do

[tired vigil]

[lucid trail]

basically

[valid jetty]

it's more powerful than that

you can bind to any stage of the match

Foo(x @ Bar(Baz(19)))

[ornate quiver]

waow

[lucid trail]

pointers!!

[hollow patio]

me when "inconsistent use of tabs and spaces in indentation" in python cuz i used nano

[lucid trail]

i have no idea why this offset is 8

it should be 4

[jade stone]

😭

I mean backport the fix from v8

[pseudo sierra]

oh

horror

[winged mantle]

why are hard things so hard

[ornate quiver]

hop on micro

[winged mantle]

learning vim is a good idea

vim bindings don't just apply to vim

[hollow patio]

if (not updateRichPresence(richPresenceProtocol, inGame, placeId, serverId, start, gameName, richPresenceDetailsFormat, gameIcon, userID)):
                richPresenceProtocol.close()
                richPresenceProtocol = richPresenceConnection()

good to go

[winged mantle]

why are you using parens in if

fake python dev

[winged mantle]

I start doing something and then give up because it seems really hard

[lucid trail]

could this be because it’s on 64bit

[hollow patio]

i found the issue

aparently i was feeding a boolean to pypresence

instead of the real stuff

gameIcon = getGameIcon(placeId)
        while (inGame and not gameIcon):
                gameIcon = getGameIcon(placeId)
                time.sleep(15)

[jade stone]

guhhhhhhhhhhh

thanks vite

(this error is from a plugin, not vite)

god forbid they tell you the plugin that errored

[lucid trail]

@valid jetty do you intentionally make the stack grow downwards? I add my offset, but in elle it has to be subtracted

[lucid trail]

I think nothing is said about the ordering of local variables on the stack, could be a QBE thing. seems like local allocations growing downwards is typical

[jade stone]

guhhh what horror code did i write where html is resolving to javascript

[lucid trail]

looks like this has something to do with how cranelift generates the assembly actually, the stack slots are merged

[nimble bone]

@jade stone satan

[jade stone]

Oh

Only one image sent here

[valid jetty]

no, however the stack grows downwards on 99% of modern cpus

i'm not sure why it grows upwards for you

also subtracting 8 from x's addr is undefined behavior

elle's order of allocations is not the order of variable declarations

[solid gazelle]

SEND THE REST THEN

[solid gazelle]

the rosie grows downwards

[solemn ravine]

@woven mesa

[solemn ravine]

@woven mesa https://github.com/Lakr233/ConfigurableKit

[woven mesa]

tyy

[lucid trail]

i would expect so

[winged mantle]

why i do i question the decisions my previous self made

they're always right

i always had a good reason to do a dumb thing

[solemn ravine]

https://cdn.discordapp.com/attachments/1436186705655889980/1438931023780909156/image.png?ex=6918ac72&is=69175af2&hm=48ad066076310d6635e1f9d390710fcbec2f102e3936149fe446a7eeb81c7e41& @woven mesa

[valid jetty]

at the very least, variables allocated within other scopes are put into the root scope

that way things are only allocated once

the behavior before was that if you declared a variable in a loop it would allocate new stack memory for it each time and cause a stack overflow

which is obviously horrible and not what you want

well qbe does that for me

https://en.wikipedia.org/wiki/Register_allocation

http://c9x.me/git/qbe.git/tree/mem.c

[lucid trail]

interesting

[valid jetty]

that's actually why i have to pass every pointer through a noop function when the gc is enabled

since i don't have access to the assembly level i can't use the registers as roots for the gc, so i need to make sure all variables are on the stack so they're discoverable as roots within the gc

[lucid trail]

this is not optimal is it

and how does passing a pointer through a noop function ensure variables are on the stack

[hearty lintel]

how did u do

[valid jetty]

i guess it means qbe can't optimize it

who knows

[valid jetty]

no, it isnt

i need to ask in the qbe forums if there is another way to do this

maybe a way to get access to the registers

or some other way to ensure variables are always put on the stack

[lucid trail]

how are you interfacing with QBE in rust rn?

just curious

wait why do i sound so much like an LLM, kinda concerning

i rarely reason

if i were to reason id probably do most of it in latent space

[hearty lintel]

@hoary sluice holy did u see the message sent by the top two teams?

[valid jetty]

i wrote my own qbe bindings that implement fmt::Display to write them into qbe IR

[winged mantle]

seems like it invokes a command

oh jumpscare

[valid jetty]

theyre based on a qbe binding library but ive added maybe an extra 5000 lines of code to them so theyre not really based on that anymore

look in this folder https://github.com/acquitelol/elle/tree/rewrite/src/compiler/qbe

this is where all of that lives

[lucid trail]

qbe IR has no way to spill it? it wouldn't make sense to not have a stack store instruction

[lucid trail]

pretty amazing

[valid jetty]

wdym stack store

they have stack alloc (with an alignment) which returns a pointer that you load and store from like any other pointer

[valid jetty]

well it invokes qbe, after turning the code into qbe IR

i think theres probably a way to wrap qbe itself into an amalgamation and use it like a library

but this is fast enough and i dont really care about the speed up from not needing to write and read from a 5k-ish line file

[lucid trail]

ok yeah i see why you'd use that trick to ensure the variables always on the stack

[valid jetty]

because if i dont, qbe will optimize naive stack allocations into registers

consider

fn foo() {
    a := #alloc(i32);
    b := a;
    a = nil;
}

b is just assigned to a pointer

its a naive stack allocation

it can be put into a register

so qbe will do that

however if you call a function, qbe decides to put it on the stack (since you allocated the variable on the stack originally using alloc8 or whatever, i assume it wants to keep invariants)

optimizations can only happen when youre sure that the code behaves exactly the same

you dont know if the function expects your pointer to be on the stack

[hearty lintel]

i wish i can post a screenshot of the results of the coding contest because this is truly crazy

[valid jetty]

since it was allocated on the stack you should pass it on the stack, but if you know it will only be used in the current function you can put it in a register

i hope that makes sense

[lucid trail]

yes and the noop function is what makes it stay on the stack

makes sense

[hoary sluice]

yes

theyre from my home location

st pölten

[valid jetty]

in my case, i traverse the stack linearly looking for things that look like pointers to see if theyre allocated via the gc

[hearty lintel]

[valid jetty]

if i have a naive case, qbe doesnt know that the variable actually should be on the stack

[hoary sluice]

terrible

[valid jetty]

since its just used in the current function

[hoary sluice]

#41 in vienna

[valid jetty]

so itll put it in a register

[hearty lintel]

same

[hearty lintel]

wait thats not bad

[hoary sluice]

it is bad

last year i was #1 in st pölten

in october and in april

[lucid trail]

at what stage does this happen? within each block?

[hoary sluice]

how did u do

did u solve 6

[hearty lintel]

AI has changed the game this time

[hearty lintel]

no i couldnt even solve 5

[hoary sluice]

we did level 4 with ai and level 5 without

3h50m leveh 5 solve

[valid jetty]

gc collection happens when the allocation capacity is full https://github.com/acquitelol/elle/blob/rewrite/std/runtime/allocators/gc.le#L180-L182

[pearl stag]

gc.le: Lines 180-182

if self.total_size + size > self.threshold {
    self.collect();
}

[hearty lintel]

🔥

[hoary sluice]

i was vibecoding for like 2h but like not blind

didnt help at all

[valid jetty]

the capacity increases if a lot of things are still alive after collection* https://github.com/acquitelol/elle/blob/rewrite/std/runtime/allocators/gc.le#L164-L176

[pearl stag]

gc.le: Lines 164-176

if before_size > 0 {
    ratio := freed_size * 100 / before_size;

    if ratio < 25 {
        self.threshold *= self.growth_factor;
    } else if ratio > 75 {
        self.threshold /= self.growth_factor;

        if self.threshold < GCAllocator::MIN_THRESHOLD {
            self.threshold = GCAllocator::MIN_THRESHOLD;
        }
    }
}

[hearty lintel]

wait really? how did the top two teams do it then?

they said they only used AI and didnt even look at the problem

[valid jetty]

eagely should i start vibecoding

[lucid trail]

wow, if ptr && ptr > 0x1000000 {, nice

not what i was expecting

[valid jetty]

this is genuinely how you check for pointers lol

though, this is fine

im just pruning the list so i dont check things unnecessarily

[hearty lintel]

i mean the top two teams just won a coding contest using nothing but AI

[valid jetty]

i still check if the value is within the range of an allocated pointer

[hearty lintel]

they were very transparent about it and even rejected the prize

[valid jetty]

wdym

what uni stuff

no i don't work

i'm just wondering what you're talkng about

oh personal statement?

yeah that's for uni applications lmao

[hearty lintel]

i couldnt send my last message because the automod thought it was spam? weird i was just saying they were very open about what they did and rejected the prize money, they just wanted to send a message that AI is killing the spirit of competition

[lucid trail]

i cant say i looked into this that much when i chose lol

[valid jetty]

well i only did that to such a deep level because i don't have amazing grades and i need to have a very good personal statement and reference in place of that

[hoary sluice]

@hearty lintel first place in vienna #5 global on awards ceremony

"did you use ai?"
"yes."
"a lot?"
"yes."
"solely?"
".. almost"

[hearty lintel]

were those different guys

[lucid trail]

maximizing for fun is good

[hoary sluice]

where did they say that

maybe they had good prompt engineering 😭

[hearty lintel]

its on the CCC discord

i cant send it here because brainrot role

[lucid trail]

math is so fun!! and less coding than cs, which i find quite tedious. theres a lot of math in cs, no?

[hearty lintel]

#1090215887862251550 message

heres a link

oh i didnt type fin-ish but maybe it was another word

ok im just gonna copy what they said here