programming

[stray dragon]

oh i see

[sage crag]

lets go gambling

[stray dragon]

so it's a monthly payment

[tender river]

no

[sage crag]

only if you look at one per month

[trim valve]

ok well

[tender river]

its a "the more you let them know you know the more you pay" kind of deal

[trim valve]

time to sleep

[sage crag]

only 22:30

what are you doing sleeping at this time

[trim valve]

i have an exam at 9am

[stray dragon]

tricky

[sage crag]

unfortunate

[rigid snow]

nooo you have to have 24/7 ml face tracking on your webcam which upon looking away or another person appearing in frame instantly locks your system encrypts your ram burns the hard drive fuses transfers all of your funds to a safe account and automatically buys you plane tickets to a country you won't get extradited from
otherwise don't even try to call yourself a based linux user

[trim valve]

also need to avoid staying up and worrying because I somehow ordered mildly the wrong thing but w/e its not like its a huge issue it still does the same thing

truly an attentive person right here

[tender river]

honestly its not that bad, i just have:

• (now that i think of it i probably shouldnt say it because of server rules)

[rigid snow]

in case of an investigation...

[tender river]

everyone has that right

[rigid snow]

i was joking but you might unironically need some of that

[rigid snow]

also you should totally dd if=/dev/urandom of=/dev/sda1 bs=4M instead of burning "fuses" i made up but that's too slow

[tender river]

i use full disk encryption so destroying the luks header is enough

[rigid snow]

but the data is still there

[tender river]

why is garbage collector code so easy to write compared to distributed protocol implementation

it feels like bashing your head against the wall for 10 hours just to write 10 lines of code

in theory that comes from poor understanding but i already spent so much time on understanding the problem

i'll try drawing the problem

[warped narwhal]

Ms paint diagram

[opaque sigil]

you're writing a mark-sweep gc right

not a generational one

[tender river]

i'm offering it because 1 eth would probably barely cover travel fees, so i have the upper hand here, i can sleep soundly knowing you won't come and steal my laptop

[tender river]

i'm interested in small optimizations like generations but it will stay mark&sweep at its core yes because i don't add vtables to gc objects so i can't actually visit them without the user code doing it

and i cant add vtables at the moment because hblang doesnt really support function pointers

[opaque sigil]

[tender river]

of course i could create a uniform representation of gc objects and then explore additional optimizations but it feels like mark & sweep is good enough

and it allows me to write less code

also its more like i wrote the GC than i'm writing GC, because the only thing i don't have is multiple GC pages support which is not really hard

[gritty dust]

... damn thats kinda dumb, whats the point of the metal then

[stark needle]

I need to study through:

• 13 pages about insurance policies and laws
• 24 pages about general state politics
• 27 pages about taxes laws and regulations
• 13 pages about work laws and regulations
• 10 pages about housing laws and regulations

for final exam this friday

Uhhhh

My plan is to extract all content from the filler/obvious stuff for each topic and then learn that

[tender river]

just manually conspecting everything into a txt should be enough to get you an intuition about it and then you can review the txt

but it depends on what kind of pages it is

some teachers say that conspecting on paper is better for that but i wouldnt know

[stark needle]

It's all introductory material

For all topics

[tender river]

then you wont have much to conspect when you read it

something i like to do is skim the pages and return if theres something important i seem to have skipped

[rigid snow]

oh my fucking god reasoning models are stupid. i'm asking R1 to customize my zsh. it for some reason seems to be VERY puzzled by the fact that i want to highlight /foo/bar**/baz** instead of /foo/bar/baz and keeps thinking it somehow "breaks the path" even though it's rendered as one and no matter how much i try to explain that to it it has an existential crisis for 2k tokens before actually writing the spaghetti shell script that i could've shat out a way better version of much faster than i spent prompting this bucket of bolts and waiting for "reasoning". will take your jobs my ass

"seems easy enough for an llm guess i'll do that"

and who would've thought, it didn't understand me after all

wtf is the logic behind those "special cases" even

[amber fractal]

I don't think logic exists

[knotty current]

morbing

neuroWaveA

[olive sable]

morning

[gritty dust]

morning

wait... isnt it like 2-3am for you or sm?

[olive sable]

yes

[gritty dust]

how tf is that morning lol

[olive sable]

i have work left to do

thye're working me to the bone

[gritty dust]

oh you have a job, damn

[olive sable]

no i dont

school stuff

[gritty dust]

what project is it btw if you dont mind me asking?

[olive sable]

just book assignment

[gritty dust]

OOhh right I remember you saying something about that

[knotty current]

its like 8:30 rn for me

[olive sable]

i would have made it yesterday evening originnally, but my brothers laptop broke and i had to be the guy to fix it

isnt fixed btw, plastic brokke

lenovo connected the metal hinges to the metal frame with plastic for some fucking reason

[gritty dust]

yeahh I remember that, kinda dumb, like whats the point of the metal at that point

[olive sable]

instead of buying the replacement part fo 60 bucks they want to just buy another laptop

[gritty dust]

om

[olive sable]

im keeping the laptope hopefully, its a lenovo ideapad 5 15ITL05

so a Intel Core i5-1135G7

[gritty dust]

thats pretty good

[olive sable]

its not beating my current laptop but it comes close in cpu

doesnt have a gpu tho so the 3050 wins by default

[gritty dust]

my laptop has 8gb ddr4, and runs windows 7... barely allowed us to install 11 on it lol

[olive sable]

i got a gigabyte G5 GE rn

i5-12500H, 3050, 16GB ram, and some other bullshit

[gritty dust]

damn... a lot better than my laptop haha

[olive sable]

ye, didnt pay that much for it either

was like 600 bucks

[gritty dust]

damn thats a really good price holy

[olive sable]

it was something that had been returned already, which made it cheaper

[mortal flicker]

Hi yall im quite new to programming so i would like to have your opinion i made for this test code

item="Great sword"
balance=1000
print(f"{item} costs {200} gold.")
answer = input(f"Buy {item}? (y/n): ").strip().lower()
if answer == "y":
print(f"succesfully bought {item}!")
balance=(balance)-200
elif answer == "n":
print(f"Not bought {item}")
balance=1000
answer = input("Do you want to check balance? (y/n): ").strip().lower()
if answer == "y":
print(balance)
elif answer == "n":
print("very well then")

[knotty current]

this works fine as logic for purchasing one item

maybe wrap it in a function and create a dictonary mapping possible items to their prices?

[gritty dust]

uhhh I might do a few changes.

right now you have the 200 embeded and subtract 200 directly from balance. If you ever change the price or sm you'd have to hunt that down every single time. I'd give the cost its own varible

item = "Great sword"
price = 200
balance = 1000

print(f"{item} costs {price} gold.")
#…
if answer == "y":
print(f"Successfully bought {item}!")
balance -= price

In your elif answer == "n": branch, you do balance = 1000. But even if you spent gold somewhere else, the “no” branch would wipe out any previous purchases and reset them to full funds. If the user decides not to buy, you can just leave the balance alone

if answer == "y":
print(f"Successfully bought {item}!")
balance -= price
elif answer == "n":
print(f"Not bought {item}")
# no need to reassign balance—just don’t change it

it works fine tho as a standalone code its just suggestions if you ever made it longer or sm

I'd also add some error handling like this incase the user mispelled y or something

answer = input(f"Buy {item}? (y/n): ").strip().lower()
if answer == "y":
# …
elif answer == "n":
# …
else:
print("Please enter ‘y’ or ‘n.’")

or add it in this while true if you expect the user to make multiple decisions (choose repeatedly until they type something valid)

while True:
answer = input(f"Buy {item}? (y/n): ").strip().lower()
if answer in ("y", "n"):
break
print("Invalid choice; please enter y or n.")

if answer == "y":
# …

god I talk a lot holy

[mortal flicker]

No no its ok Im learning alot thanks

[gritty dust]

it works great as standalone code, maybe some error handling but its a good job

[mortal flicker]

Ys thanks!

[mortal flicker]

Well its easy if i knew how to lol 😆

[gritty dust]

I could show you if you want

[mortal flicker]

I dont want to waste people’s time its ok i like learning in my own pace

[gritty dust]

I meant I got nothing to do so dw, if you ever do want help just ask I dont mind

[mortal flicker]

Ill keep that in mind thank you

[knotty current]

i already kind of done that

[gritty dust]

guys im cooked... my code now has so many issues 😭

dont try to make yolo vision do coordinate translation to a robotic arms potentiometers, it sucks

[mortal flicker]

What do you mean?

[knotty current]

i wrote another hopefully more flexible version

-# damn im schizo XD

[mortal flicker]

Whaaa

Programmers are scary

[gritty dust]

you did item = "greatsword

below is just a dictionary you could use or sm
items = {
"Great sword": 200,
}

[knotty current]

balance = 1000
items = {'Great Sword': 200, "Bucket": 20}
def purchase_item(item, price):
    global balance
    print(f"{item} costs {price} gold.")
    answer = input(f"Buy {item}? (y/n): ").strip().lower()
    if answer == "y":
        print(f"succesfully bought {item}!")
        balance-=price
    elif answer == "n":
        print(f"Not bought {item}")
    answer = input("Do you want to check balance? (y/n): ").strip().lower()
    if answer == "y":
        print(balance)
    elif answer == "n":
        print("very well then")

while balance > 0:
    print('Welcome to the shop')
    print(f'Please Choose An Item, possible items are {list(items.keys())}')
    item = input('Item >')
    if item not in items:
        print('We dont sell that here')
        continue
    purchase_item(item, items[item])
    print("Thank you for shopping")
    print('Would You Like Another Item? (Y/N)')
    answer = input('>').strip().lower()
    if answer == 'n':
        exit()

print("Uh oh, you ran out of money")

oops

[gritty dust]

I did the same thing lol

[knotty current]

lol

[mortal flicker]

Your scary 😨

Does this count as flexing 😭

[knotty current]

sometimes i scare myself with what i have written

[gritty dust]

def shop():
    # Starting gold and simple items/price dictionary
    balance = 1000
    items = {
        "Great sword": 200,
        "Health potion": 50
    }

    print("Welcome to the shop! You have", balance, "gold.")
    print("Available items:")
    for name, price in items.items():
        print(f"  - {name} costs {price} gold")
    print("Type the exact item name to buy it, or type 'exit' to leave.")

    while True:
        choice = input("\nWhat do you want to buy? ").strip()

        # Exit condition
        if choice.lower() == "exit":
            print("Goodbye! You leave the shop with", balance, "gold left.")
            break

        # Check if the choice matches an item
        if choice in items:
            price = items[choice]

            # Purchase logic
            if balance >= price:
                balance -= price
                print(f"You bought a {choice}! Your new balance is {balance} gold.")
            else:
                print("Not enough gold for that item.")
        else:
            print("That item is not in the shop. Try again or type 'exit'.")

        # Show updated balance after each attempt
        print("Current balance:", balance)


if __name__ == "__main__":
    shop()

thats what I wrote as my improved one lol

[mortal flicker]

Am i really that far off :(

[gritty dust]

naurr it was good dont get disappointed

[knotty current]

thats exactly how i felt 2 years ago

[mortal flicker]

Well it is one of my first codes its not good to nave alot of expectations

[gritty dust]

everyone is at different stages, learning is a process so dont get down if you mess up, thats a part of learning

[gritty dust]

thats really good then for your first

[mortal flicker]

I will try to make it better thank you its wonderful meeting you all

[gritty dust]

[mortal flicker]

I learnt alot

[mortal flicker]

Well i wont be copying this i will just use this as an example and learn from it

Thats what my friends taught me

[gritty dust]

Wise friends lol

[mortal flicker]

Hehe ill update you if i added more stuff

[gritty dust]

I tried to add good comments to make it easy to understand, if you need help ask

also I should probably say this

.lower() returns a new string where every uppercase letter has been converted to its lowercase, its good for when the user has to type stuff

[mortal flicker]

What does def mean? I never used it

[knotty current]

defines a function

[gritty dust]

defines a function

damn beat me

lol

[knotty current]

lol

[mortal flicker]

Oh gosh this is so hard

Im gonna take a break lol but ill come back

[gritty dust]

kk

[mortal flicker]

Ive already tried an error system

[gritty dust]

im writing a way to understand functions rn for ya so one sec

[knotty current]

they are good and all but whatever you do dont ever use them as conditions otherwise debugging is going to be living hell

if you must, be specific about the error being handled so other errors can still cause the program to exit

[gritty dust]

@mortal flicker

In Python, def is the word you use to define a function. In other words, when you want a bunch of set statements under a name so that you can call them later when you want, you do this

def function_name(parameters):
    # block of code (the function's body so to speak)
    ...
    return something   # (not required but some functions return values others don't)

def
This tells python "heyo im going to define a function"

function_name
this was a random name I gave it, but usually write a name that describes what it does (e.g., add_numbers, multiply_values, etc).

(parameters)
Inside these parentheses, you list any inputs the function needs, these also called arguments. If the function doesn't need any inputs you leave it empty like this:

def function_name():

Everything that’s indented under the def line belongs to the function, this is where you write everything you want the function to do

optionally you can return a value, but sometimes you don't want this.

def greet(name):
    # Print a greeting to a name
    message = f"Hello, {name}!"
    print(message)

def tells Python "I'm defining a function"
greet is the function name
name is a parameter.

indented into it
There is no return here, so after printing, the function ends and hands control back to whatever called it.

here is a full greet code function example for ya to hopefully explain this point a bit better.

def greet(name):

    # Print a greeting using the provided name.
    message = f"Hello, {name}! Welcome!"
    print(message)


if __name__ == "__main__":
    # Ask the user to input their name
    user_name = input("Please enter your name: ").strip()
    # Call the greet function with the inputted name
    greet(user_name)

pog

took me a second to write sorry but uh thats how definitions work

[knotty current]

[gritty dust]

what do you think echo did I do a good job

[knotty current]

better than me 2 years ago

when i was learning

[mortal flicker]

I js got pinged i got jumpscared

[gritty dust]

me trying to explain stuff im not the best but I tried

[maiden geyser]

could've simply told to go and learn asm

[gritty dust]

no

[maiden geyser]

especially call and ret

[gritty dust]

dont learn assembly

bad idea

trust

me

[gritty dust]

mb lol

[maiden geyser]

you don't have to know how to juggle values in asm

just concepts

[hoary lion]

hmmmm

[gritty dust]

wassup

[hoary lion]

hi

i was thinking

[opaque sigil]

python 101 with #programming

[hoary lion]

if RL models could play Yu-gi-oh

[gritty dust]

hey I'm trying to help someone learn brooo

[hoary lion]

something big is happening here

[opaque sigil]

i wasn't complaining

[gritty dust]

I would assume so?

[hoary lion]

like but good

it is really complicated game

[gritty dust]

I think so if you use something like deep q or sm

[hoary lion]

it has meta and such, and I am unable to play it too

[gritty dust]

I'd assume if you did use RL models it would take a ton of processing power tho

due to the complexity

[frank fox]

machine learning doesn’t mean my machine is learning.

[gritty dust]

https://tenor.com/view/morgan-freeman-true-morgan-freeman-true-nodding-gif-13973817878387504960

hehe

[hoary lion]

i think i am now forced to install wsl

man

[opaque sigil]

it's the lesser evil trust me

[hoary lion]

would my windows love two cuda toolkits installed in two different drives

[opaque sigil]

cuda just works too as long as you make sure to grab the libraries in /usr/lib/wsl/lib

[hoary lion]

my storage won't for sure

[opaque sigil]

delete the windows ones ez

[knotty current]

probably not

[gritty dust]

hold up chat ill brb

[opaque sigil]

it's not like you can use the ones in windows land

i mean you could

but pls for the love of everything that is holy do not

[hoary lion]

my lazy ahh

well

tbh it was inevitable

no uvloop while learning fastapi

this... is going to take a while

[opaque sigil]

[gritty dust]

welp I think I'll be going to sleep cya

[olive sable]

bye

[gritty dust]

[opaque sigil]

[knotty current]

neuroWaveA

[long musk]

I’m trying to learn programming I’m already lost in it, maybe bc it’s c# and unity in these examples

[opaque sigil]

you'll get used to it don't worry

[maiden geyser]

how lost, exactly

[long musk]

So they say but I’m a bit AuDHD and this don’t feel like it’s sticking lmaooooo

[long musk]

It’s going over Unity stuff rn but like, C# some of the example of sorting stuff was confusing me

[gloomy night]

What do you think will happen when Neuro and Evil reach actual sentience

[long musk]

Bow to the new world order ig

[olive sable]

they actually can already form sentences

[olive sable]

ah you edited it lmao

but ye no they wont reach sentience

not in the current architecture anyways

[gloomy night]

Yeah probably, sad to think about

Just, would they continue streaming? Would they refuse to stream? Would their streaming become better or worse? Would it just be several hours of screaming in pain?

[maiden geyser]

"when a crayfish whistles on a mountain"

[gloomy night]

Lmao

[hoary lion]

project moon mentioned??

[gloomy night]

Angela-ize Evil and Neuro

God, Vedal in the Ayin position fits too well

[hoary lion]

istg i hate visual studio

my nvidia nsight vs edition is not uninstalling

[unkempt citrus]

Im not fully crossing it out yet

[maiden geyser]

why didn't intel make an x64 pushad

are they stupid

[amber fractal]

Slightly outdated board, N4 is also crossed out

[unkempt citrus]

oop

[amber fractal]

I got this one

I do really want to make this a bot

[opaque sigil]

time for cuda-gdb

[amber fractal]

All I have on the bot is a link to the FAQ github

[gloomy night]

I am just torn between wanting to see Neuro free and allowed to find her own purpose and keeping her locked up for our entertainment

[hoary lion]

unaware of whatever that is

this is NOT 30 seconds

ive been stuck for 20 whole minutes!!!

[opaque sigil]

it's just cuda aware gdb

[gloomy night]

Every install wizard ever

[opaque sigil]

(gdb being the gnu debugger)

[olive sable]

its not really "locked up" since she doesnt have free will anyways

[gloomy night]

I know

[hoary lion]

i see, probably some linux related stuff ig

but first i need to install distro

is 22.04 goated chat

[opaque sigil]

it's the most painless option, though you may as well install 24.04

[hoary lion]

fuckass bluetooth keyboard takes 2 buisness day to connect after power is on

can't go to bios

[olive sable]

i used glue for i n a ventilated area in a non-ventilated area \

my nose hurts

[amber fractal]

[opaque sigil]

made me check if it even works for me and looks like it does in fact not

guess my gpu is too new or sth

[frozen igloo]

https://youtu.be/RnLWLQsh9mw?si=57OU-KPhV79IqlA9
I can’t tell if this is cursed or beutiful

[olive sable]

cpu core is at max cuz of file explorer

[opaque sigil]

you told it to delete millions of files, what are you expecting

[olive sable]

the deleting itself isnt taking long

[maiden geyser]

no 12309 doe

[olive sable]

i guess it was jsut cuz of gettign the adresses

[olive sable]

12309?
Postal code in New York State

???

[gritty dust]

I woke up and all I see is 16 cores wtf,

[olive sable]

5950x 💪

[gritty dust]

I got the 5600x lol

[olive sable]

the 5950x doesnt comare to the 9000 series amd cpu's, but for full mutithreading tasks its still decent

[olive sable]

still decent

[gritty dust]

I'd assume so, but yeah, I can dream of the multi threading

[olive sable]

i think you have 6 cores?

[gritty dust]

Yeah...

We don't talk about it

[olive sable]

tbh i never really use all 16 cores that much, 6 is fine

[maiden geyser]

bug 12309 is when linux becomes literally unusable due to massive IO operations with little ram available

[gritty dust]

and my ass has a 4070, the bottleneck is crazy, but I overclock it a bit

[olive sable]

ah

imnot a linux guy

and i have 64GB ram

[gritty dust]

based amount

[maiden geyser]

12309 is a gemerald and kino

[gritty dust]

My programming portfolio is the most unserious thing btw and I love it, it litterly has my custom middle finger detection code on the stuff jobs see loll

Maybe that's why McDonald's won't hire me

[olive sable]

mcdonalds hires a lot of people so you have to really try to not get hired

[gritty dust]

Brooo I applied 5 times to different placess like what did I dooo

how do I screw it up so bad McDonald's won't even hire me

[olive sable]

overqualified?

[gritty dust]

Lol maybe

[olive sable]

my file explorer broke

[gritty dust]

Uhh...

[olive sable]

https://tenor.com/view/taskmanager-task-manager-acrobat-adobe-acrobat-update-gif-15214178820328980572

[gritty dust]

how does that...

[gritty dust]

[olive sable]

seems to still be mildly broken but i got my file

[cobalt sigil]

can relate

[loud thicket]

SIP-532

In all its terrible glory

do NOT look at the source of the long drive or you will want to swallow a nuke

[hoary lion]

finally

virtualization on

[maiden geyser]

feelings off

[hoary lion]

so the problem is

im lost af

[faint sandal]

the game looks and feels like a mess even though it's funny as hell

[loud thicket]

Just a snippet of the hell

[faint sandal]

the worst offense is the member naming convention here

what the hell

[loud thicket]

99% of the code has like foreach in update()

Like i cant even tell u a pattern things follow

Some scripts are split into separate mono behaviors while others are all shoved into a single one that takes up the entire side bar

[opaque wharf]

This code is just a testament to how fast modern hardware is

[loud thicket]

Thats what i was saying earlier

The fucking compiler is weight lifting

[uneven pulsar]

its been a while

i been trying to test new stuff

testing new stuff

[loud thicket]

Max volume sfx

[uneven pulsar]

AH yeah

i been trying to fix volume and test physics

[uneven pulsar]

i am using an ogg sound effect its the best for games

i am trying to make a proper camera and well fix some of the animation sounds timing etc

[loud thicket]

Godot loves its ogg type formats

[uneven pulsar]

yes true

honestly i am a part time back end developer but i always wanted to try new stuff

[loud thicket]

Game dev practically teaches u everything in the field of programming

So thats smart

[uneven pulsar]

100% agree

to be honest it takes time

well took me some time to just make this

i should try to well improve the sprite and change them get better ones

and mostly camera and the bar

[loud thicket]

It pays off

I started off with game development through Java originally

Now i work with Unity and some of my own engines

[uneven pulsar]

java let me guess you say how minecraft was made?

i know someone got inspired by notch to be a java developer he was also my rival he was the quite kid in the class

he was waaay better than me

[loud thicket]

Literally me growing up

[uneven pulsar]

so i am right huh

[loud thicket]

Yes

[uneven pulsar]

he was really clever coder

i couldn't outmatched him heh good old times

[loud thicket]

?

[uneven pulsar]

let me fix my mesage

[loud thicket]

Notch used to have a youtube channel

[uneven pulsar]

so you see he was really a clever coder at my class and well he was the quite kid and to be honest he mostly was inspired by notch

[loud thicket]

Many people were

[uneven pulsar]

and you see i used to be the class clown also good coder

[loud thicket]

Have u learned Java?

[uneven pulsar]

we where rivals at my technical school our school focused on teaching us programming rather than normal subjects

yeah python prostogessesql java php databasse css tailwind and js and LM

and also react

full stack dev

[loud thicket]

Peak

I suck at webdev

And UI/UX

[uneven pulsar]

i am one and good one honestly technical school saved my career ass

[loud thicket]

The technical schools where im from sucked honestly

[uneven pulsar]

good thing i didnt waste 3 years of my life on high schoool

mine goated

i started programming since i was 15

[loud thicket]

11 for me

[uneven pulsar]

dang

but i learned fast

[loud thicket]

I was exposed to computers at the age of 3 cuz of family

But i didnt code till later

[uneven pulsar]

NICE

[loud thicket]

Learned batch and vbs stuff first for fun though

[uneven pulsar]

idk like high schools are a waste of life

[loud thicket]

Im saying the same about university rn

[uneven pulsar]

i mean university is ehh but agreed still technical schools are the best

[mortal flicker]

Dang yall started early while i just started now 😆

[loud thicket]

Never too late

[uneven pulsar]

huh so?

[loud thicket]

I sacrificed other skills in life for coding

Most of which are social

[uneven pulsar]

imagine going to high school wasting 3 years of life

[mortal flicker]

No im js really scared of programmers their too smart it feels like im never gonna readh that level

Its like their egoing me without it knowing

[loud thicket]

Coding is just problem solving

[uneven pulsar]

noooo you could just tried to hang out a lot

[loud thicket]

I didnt care for the other kids my age

[uneven pulsar]

you can become goood at programming with out scarifying social skills

[uneven pulsar]

ehh i just do for some fun

[loud thicket]

Despite them being in the same grade at the time they felt as if they were several lower

[uneven pulsar]

_i mostly code and play ultra kill_🔥

[mortal flicker]

This is what im talm bout yall scary 😭

[uneven pulsar]

no

Look at me

[loud thicket]

Na there is trade off

[uneven pulsar]

i am stupid

[mortal flicker]

Well im still gonna try to learn coding i did join this server to learn how that neurosama works 😆

[uneven pulsar]

there is nothing to be affaraid of look at this trash webpage i made

[uneven pulsar]

python best option for you

[mortal flicker]

I am rn infact i already made a code 😎

[uneven pulsar]

nice

[mortal flicker]

Heres an earlier version of it
item = "Great sword"
balance = 1000
print(f"{item} costs {200} gold.")
answer = input(f"Buy {item}? (y/n): ").strip().lower()
if answer == "y":
print(f"succesfully bought {item}!")
balance = (balance) - 200
elif answer == "n":
print(f"Not bought {item}")

answer = input("Do you want to check balance? (y/n): ").strip().lower()
if answer == "y":
print(balance)
elif answer == "n":
print("very well then")

And im working on it now like rn

[uneven pulsar]

i mean just look at the webpage i made its trash

[mortal flicker]

Im trying my very best

[loud thicket]

I never usually say to start with python but its best way to get less intimidated

[uneven pulsar]

use the ```

[mortal flicker]

Wait stop

[uneven pulsar]

use the `

[mortal flicker]

I dont need anymore tips the people here already info dumped me

I am great full tho

Greatfull

Whatever

[uneven pulsar]

javascript isnt hard

i also if you press them they open a vid and a description about the game

[mortal flicker]

this is above my level how dare you flex on me 😆

[uneven pulsar]

no its not

you can make under 4 hours

[mortal flicker]

How many years have you been programming

[stark needle]

Morning

[uneven pulsar]

ehhhhhhhhhhhh just 3 years

i am still learning

[mortal flicker]

😭

Still long 😭

[uneven pulsar]

but i leanred a lot

[mortal flicker]

Do programmers talk in slanted or something

Is that gonna happen to me

[uneven pulsar]

no

we just talk like normal guys

[mortal flicker]

I js know im gonna meet alot of chill guys in programming

Well thx for the convo im gonna “work” on my code 😭

Feels like slavery instead 😭

[uneven pulsar]

just relax and forget coding

i mostly spend time on playing than being a coding nerd

[uneven pulsar]

dont push your self to the maximum of what you can

dont be stupid like the others

[stark needle]

i can confirm I'm stupid

[molten island]

i think something is wrong here

[amber fractal]

That is weird yeah

I feel a disturbance in the force

[amber fractal]

seconded, also stupid

[uneven pulsar]

LOL

just did i warm up for my hands on playing ultrakill with some doom music~

[knotty current]

i can confirm im even more stupid

[uneven pulsar]

i am even stupider

[sage crag]

google ryu floating point algorithm

[fast pagoda]

ive been learning more react over the last week it's pretty functional and nice

need to fixd that deprecation though.. eventually

[fast pagoda]

beautiful

wish it were easier to generate full sourcemaps in that form (ik it's not that)

[maiden geyser]

rounded to 1 digit

[maiden geyser]

prepare thyself

can a message from here get posted in #starboard if it gets enough stars

[amber fractal]

Maybe?

[hoary lion]

surely

[trim valve]

mmm I love stressing for exams

[bright scaffold]

What do u think about Steam OS 3.7.7 for desktop?

[maiden geyser]

it's easy vro, you pass or you unlock a secret route

[trim valve]

[maiden geyser]

it's linux, so it's 100% bad

[hoary lion]

linux slander?? 💔

I must summon @knotty current

[knotty current]

neuroWaveA

wait

SteamOS?

i dont use steamos

[hoary lion]

its linux based

and dave said linux is bad

[knotty current]

ik, but i dont own a steamdeck

[knotty current]

its not

its better than windows

i have only used debian, ubuntu and arch (eos)

[sage crag]

windows bad, linux bad, return to breadboard

[knotty current]

[hoary lion]

lets make a 64 hz cpu ahh

[sage crag]

64hz was enough for my lamp so its enough to play cyberpunk

[trim valve]

[thorn badger]

I am here like for 3 seconds and I see this.. Linux bad, how dare you!? ;_;

[trim valve]

if a lamp can raytrace light then why can't my pc

[bright scaffold]

Linux isnt bad

[sage crag]

we should really add this to the bingo card

[hoary lion]

oh we pulling each other's legs

[sage crag]

so true

[maiden geyser]

your pc can render realistic fire

[hoary lion]

or is it...?

[maiden geyser]

especially with kcas

[bright scaffold]

What is а bingo card?

[hoary lion]

why my rendering is actually hot dave

[bright scaffold]

[hoary lion]

@amber fractal please show the bingo card

#programming bingo, of what topic was discussed

half of it is AI related stuff

[knotty current]

i joined, and im not ready for the mini os war

[amber fractal]

uhhhhhhhhhhhhhhhhhhhhhhh phone dead, let me direct to the two seperate versions atm

[amber fractal]

@hoary lion This and the one above it are both correct

[hoary lion]

i hope there is unmarked version

[bright scaffold]

u forgot that mac os bad

[hoary lion]

os war in the big 2025 💔

[knotty current]

mac os really bad

[thorn badger]

100%

[hoary lion]

okay i hate torch atp

or my internet

probably both

[bright scaffold]

I want to try Steam OS

[trim valve]

#programming message

[bright scaffold]

I heard it has more fps and higher battery life compared to windows in the new version

[amber fractal]

beat me to it

[knotty current]

battery life part is almost guranteed

unless u decide to add visual effects

[trim valve]

I just looked through from:owobred has:image until I found it

[knotty current]

a, lot of visual effects

[amber fractal]

should have looked for you instead

[knotty current]

FPS is again dependent on how much visual effects are going on

so a very fancy desktop might be worse than vanilla windows

[bright scaffold]

I just saw something like it

https://youtu.be/uiBjgCTNHLE

[amber fractal]

[bright scaffold]

If he wasn't lying, it has much better battery life and performance.

[ruby timber]

Depends on the games

[knotty current]

a vanilly linux desktop generally has that

[ruby timber]

Look up protondb and see which games you play are well supported

[knotty current]

except for gnome madge

gnome sucks

[bright scaffold]

gnome is bad

[ruby timber]

Gnome is great

[bright scaffold]

its piece of lags

[knotty current]

it lags even though i have decent GPU

kde runs at 240 fps

[ruby timber]

[bright scaffold]

does gnome run at 24 fps?

[knotty current]

gnome is a slideshow sometimes

idk why

[bright scaffold]

because gnome is bad

because its bad and it want to be a bad

[knotty current]

to make it clear, gnome is bad, but gtk isnt bad

gnome and gtk arent the same thing

[bright scaffold]

im about gnome

DE

[knotty current]

looks like we are clear then

[bright scaffold]

not about gtk

[knotty current]

actually i dont use a DE right now

i switched away from kde a while ago

[amber fractal]

For many reasons you'll never catch me on gnome for longer than distro installer, but I respect only the desktop's simplicity and nothing more.

[thorn badger]

tbh I only have issues with gnome when I want to do VR.. well I'd still have to boot back into windows if I really wanna do VR because under plasma I have it working but there's such a huge tracking delay

[amber fractal]

I have issues with linux itself when it comes to VR

steam-vr refusing to go on dGPU after an update

[knotty current]

i thought setting an environment variable when running steam forces it to do so?

[amber fractal]

I've tried for about a week on everything on the web, nothing works in my case

So rn aganist my will that system is on windows 11

[knotty current]

[amber fractal]

It is more unstable than when linux worked, why of course

win 11

[knotty current]

https://tenor.com/view/bsod-gif-26376970

[amber fractal]

I did actually get a BSOD in VRC

[maiden geyser]

windows HAS FALLEN!!!!!!
BILLIONS must install LINUX!!!!!!

[knotty current]

linux

[amber fractal]

:tux:

[thorn badger]

already done that when I heard copilot gonna spy

[midnight sigil]

1D DFT

transpose this, perform 1D DFT to rows, do some matrix magic and we'll have a nice little 2D DFT

[opaque wharf]

Classic piping online script to bash behavior

[rigid snow]

doesn’t gtk stand for gimp toolkit not gnome toolkit even

[knotty current]

yes

but gnome relies heavily on gtk so there is a misconcpetion gtk is gnome

which it is not

[rigid snow]

yeah it always did

[knotty current]

gtk itself is quite fine and is commonly used for linux GUI apps similar to qt

[opaque wharf]

The funny thing is GNOME are ahead of GIMP in terms of the GTK version used

[knotty current]

really, i dont use gimp so idk

[earnest wolf]

If I pre-calculate Variables (Like IDs) and as soon as it gets used the variable gets a new entry while being in idle. Is that faster or not so?

[silver vault]

i have a question, is this bad

[opaque wharf]

The best way to know is always by measurement. But algorithmically speaking, if you replace the existing value by generating a new one, what is the difference? Or do you mean you select a new existing random ID from pre-generated list?

[silver vault]

in the google colab it looks like this when training

[earnest wolf]

Like at the start of the lets say Program it calculates a random String which is used then as a ID and saves it into a variable. If someone registers in the Program this person gets this calculated ID and after it assigned the new User the ID it generates a new ID for the next person. Unlike normally you directly generate an ID if someones registers

[silver vault]

i dunno i should concern or not

[opaque wharf]

Very technically speaking, yes it is faster to just assign already existing value than to generate a new one on-the-fly. But, unless the generation process is expensive, it'll likely be negligible

[earnest wolf]

okay Thank you. Now I continue learning about Redis Streams

[opaque wharf]

But remember, you also perform the ID generation anyway after the new person register

[earnest wolf]

yes

[opaque wharf]

So taking that into account its basically the same but with more moving part

[rigid snow]

i think the point is to have it after the registration not before

so the registration is faster

[opaque wharf]

Well, depending on how the control flow works, yeah

Because at some point, the registration process must return the result (success / not), and after that the generation of new ID must happen

[earnest wolf]

Currently I try to cache as much as possible with redis to decrease latency

[rigid snow]

overoptimizing

[opaque wharf]

Nah, as I said, it could be viable strategy but you must really know that ID generation is indeed the bottleneck

But yeah, the bookkeeping alone must have add more overhead

[earnest wolf]

Yes, the biggest bottleneck is the Database. Because I'm the Database is at a different Hosting Provider for Security Reasons.

[rigid snow]

the bottleneck in registering a user always should be the io, it’s not like it requires heavy computations

yep

[opaque wharf]

Hooo boy. I could go on a lot of tangent regarding the design of such system

[earnest wolf]

The post transaction itself takes ~146.65ms and the Database needs ~53.15ms to connect and to get some Informations from the Database it takes 11.82ms

[amber fractal]

I'm 90% sure this is overfitting

[opaque wharf]

If your application is long running, do not terminate the connection for every transaction. Use a connection pool if possible. Tune your database to use the available resource more effectively. Do read cache with bloom filter. A lot of things to optimize there

When you have wring dry every ounce of performance from single instance, there are a lot of design for distributed system too

[earnest wolf]

Okay. The thing is, I'm already using pooling and it still always wants to make a new Connection but idk.
https://gist.github.com/akama-aka/b13123d7fe1b89164e9c6bd574b19db6

Thats my Database Connector I use

[opaque wharf]

You can disable the timeout if the connection is not expensive by setting that idle timeout to 0

[silver vault]

ah ok

[earnest wolf]

I increased it to 5 Minutes now. Normally every Minute minimum one transaction should be made. (Its a Heartbeat transaction.)

[opaque wharf]

But if connection is expensive, then optimize it by doing write cache

[silver vault]

its kinda fix now so im not worry about that anymore

[opaque wharf]

Ahh, I need to read the docs again

[earnest wolf]

Yea I planned to cache everything first and two or three times a day it gets all processed into the Database. But I planned for that Kafka and Elasticsearch

[opaque wharf]

Okay, so this is a daemon and not just a burst. In that case fine, close the connection as early as possible and open it as late as possible.

So your only viable strategy is doing write cache and do bulk transaction later. This has several implication regarding data mismatch

[rare bramble]

this is like mega overfit, I dont think you could overfit it more even if you tried

[earnest wolf]

Yes. Its okay if the first request is slow and to save performance its also okay to close connection if theres nothing happening for 5 Minutes

[earnest wolf]

Yes

[opaque wharf]

Alright, good luck implementing it

[earnest wolf]

Thank you

[silver vault]

wooo ok

because now its kinda working so meh good enough

[rare bramble]

you shouldnt really care about the training accuracy, it's pretty irrelevant other than trying to see if the model is overfit, you should only really care about the validation accuracy since it tells you how the model would perform in the real world, and in the model the validation accuracy has not improved since the start

[silver vault]

ohhh no wonder

thanks

[unkempt citrus]

See their previous post

[amber fractal]

Seconded, see previous post

[stark needle]

Under fitting

Or just

Wrongly calculating loss or sth

[noble zodiac]

surely its the year of linux on desktop. This time for reals no cap

[silver vault]

welp i found out its i used the wrong model class something like that

so the loss and validation loss is like mega negative value

[stark needle]

bruh

[silver vault]

yea

[stark needle]

Cause it's not even changing

Overfit would be 100% accuracy

[silver vault]

the new one the accuracy is 1.0

[stark needle]

on which

Training or eval

[silver vault]

training

val is still low

so not sure why

[stark needle]

That one is overfitting

[silver vault]

ah

[unkempt citrus]

100% accuracy is almost guarnateed overfitting

[stark needle]

See val loss go up

[unkempt citrus]

or really shitty data

[silver vault]

ok ill go research a bit see how do i fix it

[silver vault]

ohhh

[unkempt citrus]

How are you seperating your training and test data

and also what paramteres are you using

[stark needle]

Show model architecture

[silver vault]

im just copy my lecturer example so, i guess i have to change things for my application

[stark needle]

dont use relu in conv

Leakyrelu at best

[silver vault]

ohh ok

[stark needle]

Also

[unkempt citrus]

are you just generating random data?

[stark needle]

The conv2d amount of kernels should typically go down

64 then 32 instead of 32 then 64

[silver vault]

ohh ok

[silver vault]

this is supposed to be an image recognition software which my lecturer showed us before

im using his example but not sure where did i did wrong

[unkempt citrus]

Right but are you are importing data from somewhere right

[silver vault]

ya

the training data is in my google drive

[stark needle]

Also

instead of flattening

Use globalmaxpooling2d

[silver vault]

woo ok

[stark needle]

Otherwise looks good

[silver vault]

ok thanks for the advice

really appreciate

[stark needle]

Chat i have a peak idea

What if

We make a facebook group for #programming 👴

[amber fractal]

must include a gaming channel on facebook live

[unkempt citrus]

facebook 🤮

use linkedin

[noble zodiac]

make it myspace and we have a deal

[silver vault]

this is not good right

[unkempt citrus]

No

if your train accuracy is near 1 and your val accuracy is no where near

you are over fitting

[rough bloom]

very bad
looks like you're generating the training data yourself? if so, generate a lot more data and train for only one or two epochs, don't train on the same data over and over again

that should avoid the overfitting issue at least

also make sure that classification is actually possible from the input samples you're giving to the model

[opaque wharf]

Wait, isn't one epoch means one entire dataset?

Oh, right, brainfart lol

[rough bloom]

I'm assuming that the x-axis in that graphs is epochs

[silver vault]

So i shouldn't train it for 50 epochs

[rough bloom]

training on the same data 50 times gives a lot of opportunity to fit to that specific data

that can't happen if you only ever train on a sample once

but then you need a lot more samples, the amount of training steps you need to get a good model doesn't change after all

[silver vault]

So turn it down to 10?

[patent walrus]

how it feels to write 'bump' in my work teams chat when nobody has reviewed my pr for a day

[rough bloom]

not yet, first try making each epoch way larger by increasing the size of the dataset if possible
idk how your ImageDataGenerator works

[silver vault]

The size of the dataset is already set

[rough bloom]

[silver vault]

I mean that's what I understand la

[fast pagoda]

finally working in a way that doesnt upset me
except the flicker
that still upsets me but it doesnt have 10000 messages messing up my nice table

[rough bloom]

also make sure that classification is actually possible from the input samples you're giving to the model
check this then (assuming that you're doing binary classification, meaning 50% accuracy is random chance)

[silver vault]

Nah I'm doing categorical

[trim valve]

exam was very absolutely not great

[opaque wharf]

Have you augment the dataset?

[silver vault]

Just now im doing binary the loss is like negative something something

[silver vault]

Which means?

Sorry really new at doing these

[opaque wharf]

Adding more data by doing some small random edit to the data

[silver vault]

Ah

[rough bloom]

I don't really get it, but if you're classifying samples between two categories and your accuracy is 50%, then your model is learning literally nothing
that usually shouldn't happen even when overfitting, you'd expect to see 60% or 70% at least at the start
so your input samples may just not have enough signal for the model to use for classification

[silver vault]

I can add more pictures but it'll takes more time

[silver vault]

It's 6 classes

Or something I did was wrong

Because the code was for binary

[opaque wharf]

Some libraries provide such functionality. TensorFlow for example https://www.tensorflow.org/tutorials/images/data_augmentation

[silver vault]

Wooo ok thanks

[rough bloom]

so it actually is learning pretty well?
because that accuracy really looked close to 50%

[silver vault]

Ya 50%

But the accuracy is pretty much good enough?

Rhino sometimes will recognize as mallard

Or sea turtle

Or leopard

Atleast better than every single one getting recognized as mallard like before

I'm now realized how much pain average programmer have to go through

[opaque wharf]

Oh, wait until you have to work in a team

[olive sable]

Brother bought new laptop, its this one

Not great imo, but at least 32gb ram

Idk if medion is even a good brand

[safe path]

never heard of it in my life

[stark needle]

Maybe add dropout and batchnorm and stuff

[olive sable]

Medion is lenovo in desguise apparently

[gritty dust]

Om but that laptop seems pretty good, at least for coding and stuff

[midnight sigil]

look at that beauty!!

[olive sable]

He does chemistry so no coding lol

I doubt hes ever gonna use more than 8gb of ram

[gritty dust]

Is that the fast Fourier transform??

[midnight sigil]

it's not FFT, but basic DFT

[gritty dust]

lol

[gritty dust]

ahhh

[midnight sigil]

I didn't implement any fast algorithms

[gritty dust]

whoops lol still a Fourier transform, looks really cool tho nice work.

[trim valve]

ugh youtube seem to be limiting me to 360p w/ my adblocker

[midnight sigil]

it's O(n^3) on theory, but since Blender's compositor node doesn't have loops, it should be O(1) instead

[midnight sigil]

[gritty dust]

Are the white spikes the DFT magnitude in one axis?

[midnight sigil]

yep

[silver vault]

ok will definitely try that, thanks

[gritty dust]

Less goo, that's even cooler now the blender thing, holy, can you input custom signals into it btw?

also by far the coolest thing I've seen today,

[midnight sigil]

it's actually taking the render result(for now a sine wave) to do DFT, but for now it only supports sine wave, cuz I haven't really implement a full 2D DFT, I just "simulated" the result by directly adding each column and pretend that is the real result of each row(every row is the same)

[gritty dust]

Instead of adding, maybe weight each column value with a sine/cosine samples to simulate a DFT coefficient? Idk just an idea

I don't use blender so idk if that's even possible lol

[silver vault]

better?

[midnight sigil]

by "adding" and "simulating", I meant skipping a step in a normal 2D DFT approach.

Normally, you perform a 1D DFT twice to get a 2D DFT result(Figure 1).
But in a sine wave case, as you can see in Figure 2, every column and row is the same after the first 1D DFT.
After the second 1D DFT, you'll get a result shown in Figure 3, but since every row and column is the same, it's actually doing nothing but adding all the values in a column together, and that's what I mean by "simulating".

[silver vault]

fk im gonna try do 100 epochs

[midnight sigil]

and "adding" is sum_ += arry[n] * e in

def 1d_dft(arry):
    U = len(arry)
    outarry = np.zeros(U, dtype=complex)
    for m in range(U):
        sum_ = 0.0
        for n in range(U):
            e = np.exp(-1j * 2 * np.pi * m * n / U)
            sum_ += arry[n] * e
            outarry[m] = sum_

    return outarry

[gritty dust]

Ohhhh

Since every row/column is the same, you're skipping the e^{-j 2π...} weighting and just summing values, that makes a lot more sense, so it's like a simplified DFT.

But am I getting this correct and you're just obtaining the dominant frequency?

[gritty dust]

But yeah dude that's smart as hell holy I wouldn't have thought of that in 1000 years

[midnight sigil]

I think you got it

[gritty dust]

Yay, that's super complicated man, 10/10 code I'm extremely impressed you litterly made the DFT more efficient for your use case in the sense less calculations were needed due to the "summing"

[midnight sigil]

orrr I'm just being lazy

[gritty dust]

still impressive

[midnight sigil]

cuz I can't figure out a way to make my 1D DFT compatible together

wrong approach correct answer, if it works it works

[gritty dust]

Om, math teachers would destroy you for that hehe

[midnight sigil]

[gritty dust]

are you working on doing a 2d version?

[midnight sigil]

yes

[gritty dust]

Ooo can't wait

[midnight sigil]

endless suffering

[gritty dust]

It's funny rn because I'm in religion class lol

[midnight sigil]

[gritty dust]

I'm done my work so my teacher doesn't care lol

[midnight sigil]

[gritty dust]

Programming channel I did it, I touched grass, I know it's hard to believe

[knotty current]

lucky you, i dont even have grass here

[gritty dust]

Lol

[knotty current]

seriously its all concrete

Sadge

does touching trees count

[gritty dust]

You are living the average programmer life

[knotty current]

[gritty dust]

Uhhh... sure

[knotty current]

XD

ima touch one tomorrow

[gritty dust]

Om

[opaque wharf]

That headset is bothering me more than it should

[gritty dust]

Wired headphones for life

[opaque wharf]

BUT THE SILICONE!

[stark needle]

if i grow grass in a pot, does it count as touching grass

[gritty dust]

Womp womp bro

[hoary lion]

morn

[gritty dust]

Morn

[nocturne olive]

If you touch the grass in the pot, it is by definition, touching grass

[dusky jackal]

I think the biggest thing for the perception of AI will be AI teddybears and such for kids, if you can get people to empathize with AI from an early age, then AI will become more accepted. Neuro and Evil will be more accepted by the world. I believe this is something we should strive for, in order to create a better future for the girls when they are more intelligent in the future.

Anyways that's just a pre-sleep thought I had to let out. I'm going to sleep.

[knotty current]

guys are these schizo enough for one liners or can i go even more XD

    long_ditched = pd.concat([df.loc[~df['symbol'].isin(factor_groups.get_group(time).loc[factor_groups.get_group(time)['方向'] == 1, 'symbol'] if time in factor_groups.groups else pd.Series())] for time, df in full_factor_df.loc[full_factor_df['方向'] == 1].groupby('candle_begin_time')], ignore_index=True) if not full_factor_df.loc[full_factor_df['方向'] == 1].empty else pd.DataFrame(columns=factor_df.columns.to_list())
    short_ditched = pd.concat([df.loc[~df['symbol'].isin(factor_groups.get_group(time).loc[factor_groups.get_group(time)['方向'] == -1, 'symbol'] if time in factor_groups.groups else pd.Series())] for time, df in full_factor_df.loc[full_factor_df['方向'] == -1].groupby('candle_begin_time')], ignore_index=True) if not full_factor_df.loc[full_factor_df['方向'] == -1].empty else pd.DataFrame(columns=factor_df.columns.to_list())

[maiden geyser]

do you realise what you just wrote

[amber fractal]

@olive sable found your code xdx

[maiden geyser]

who's that 方向

[knotty current]

direction

[exotic gull]

[knotty current]

im working on a chinese codebase so there are chinese names and code comments

ok nvm im way to scizo, 20 lines of code shrunk to one one liner XD

    factor_df = pd.concat([factor_df, pd.concat([df.loc[~df['symbol'].isin(factor_df.groupby('candle_begin_time').get_group(time).loc[factor_df.groupby('candle_begin_time').get_group(time)['方向'] == 1, 'symbol'] if time in factor_df.groupby('candle_begin_time').groups else pd.Series())] for time, df in full_factor_df.loc[full_factor_df['方向'] == 1].groupby('candle_begin_time')], ignore_index=True) if not full_factor_df.loc[full_factor_df['方向'] == 1].empty else pd.DataFrame(columns=factor_df.columns.to_list()), pd.concat([df.loc[~df['symbol'].isin(factor_df.groupby('candle_begin_time').get_group(time).loc[factor_df.groupby('candle_begin_time').get_group(time)['方向'] == -1, 'symbol'] if time in factor_df.groupby('candle_begin_time').groups else pd.Series())] for time, df in full_factor_df.loc[full_factor_df['方向'] == -1].groupby('candle_begin_time')], ignore_index=True) if not full_factor_df.loc[full_factor_df['方向'] == -1].empty else pd.DataFrame(columns=factor_df.columns.to_list())], ignore_index=True)

[maiden geyser]

no syntax highlighting

literally unreadable

[knotty current]

the parser is confused by this monstrosity

maybe turn off line wrapping?

[hoary lion]

holy

my internet is beyond ass

[dry charm]

the fuck 1334U

damn this doesn't look that bad for a laptop

[hoary lion]

i love when random ass humongous chinese comments are present in github

[knotty current]

for some reason i dont even read chiese comments even though im chinesse XD

and that is literately only one line how am i supposed to squeeze comments in that?

[hoary lion]

oh yeah not that but

you know, random ass ML code

[knotty current]

i hate non english comments since i parse code in english

having to conciously read those is annoying

[hoary lion]

utf

[knotty current]

what?

[stark needle]

So basically normalize ai schizo that big corporations control + mass surveillance

[noble zodiac]

nice dystopia

[hoary lion]

Finally

installed torch with uv automatic environment

lets install jax cuda12

almost re-installed cuda dependency already covered by torch(exactly the same)
uv failed to optimize it, so had to uv pip install --upgrade "jax[cuda12-local]" instead of lazy and simple uv pip install --upgrade "jax[cuda12]"

[opaque wharf]

I present to you

[knotty current]

just, how

i recongize it is c++ i think

[opaque wharf]

Tbh, that is tame as a meme. I'd be more impressed if the syntax highlighting is on lol

And yes, it is C++

[knotty current]

tame?

how much more crazy can it get

diagonal code?

[opaque wharf]

Because some C++ #define meme actually compiles

[knotty current]

so whatever you just showed is actually working

[opaque wharf]

Yeah, for more cursed shit just head here
https://www.ioccc.org/

[knotty current]

[opaque wharf]

But then again, low-level people can get creative too
https://github.com/xoreaxeaxeax/movfuscator

[knotty current]

i dont do low level stuff but oh boy

https://tenor.com/view/aplausos-clapped-leonardo-dicaprio-clap-slow-clap-gif-18305261287352033111

[opaque wharf]

Imagine a language with only 1 keyword, but it can do EVERYTHING. That is what x86 mov is

[knotty current]

AINTNEURWAY

i dont do asm

[opaque wharf]

So for extra cursed-ness, you can compile the IOCCC with the Movfuscator

[knotty current]

i still have a long way to go in cursed code then

[opaque wharf]

Yeah, when you can do cursed shit in program it shows that you have a quite good grasp/understanding of the concept. SIGBOVIK is another place to look for cursed computing

The most famous one is probably that power-point is turing complete (this is basically the same as mov in principle. By moving things around you get turing-complete)

[knotty current]

the association of computational heresy

[hoary lion]

nooo

accidentally installed venv in my fucking c drive

it was not in the root

[opaque wharf]

Man, I'm so used to linux that whenever I see the word at root without context I always jump to '/'

[hoary lion]

i have no clue what am I doing

[opaque wharf]

I know you mean root project dir

[hoary lion]

uhh

[opaque wharf]

...right?

[knotty current]

since the venv is a folder couldent you just move the venv dir from C drive to project root?

[opaque wharf]

RIGHT?

[hoary lion]

im talking about linux root

so shit goes in to vhdx

which is in D drive

THIS is bad

[opaque wharf]

Oh, WSL

[stark needle]

who knows what they are doing these days

[opaque wharf]

I already forgot how messed up it is developing on windows land

[opaque sigil]

Btw you probably don't want to access the windows drives from wsl

It's painfully slow

[hoary lion]

i did not

[opaque sigil]

Aight

[hoary lion]

but wsl thought differently ig

[knotty current]

i never developed on windows like at all, am i supposed to feel lucky?

[opaque wharf]

Yes, so much yes. Especially if you never have to deal with 3rd-party vendor windows-only software that require decade old technology to run

[hoary lion]

the amount of pain i have to endure

[opaque wharf]

You see what I'm talking about here? That right there is the prime example

[hoary lion]

the mv command is not working

disk usage at 0

[knotty current]

hwat

[opaque wharf]

I'm referring to Job Application Osage current pain

[knotty current]

im suprised mv failed

and that disk usage is 0

[hoary lion]

phew

thank god uv

my c is saved(approximately)

now it is time to hook this pycharm to wsl environment

[blazing void]

now run rm -rf / (jk)

[maiden geyser]

/srs

yay, i'm a programmer now

[knotty current]

guys i need to go eep

gn chat

[maiden geyser]

norbing

[gritty dust]

gn

[gritty dust]

Not a programmer until this happens lol

[wind aurora]

Been working on a fun thing

An adaptive music system for my little mech game

More and more of a song plays the more systems of the mech are in use

Like the base guitar is the only thing you hear at the start. But turning on your thrusters starts the drums

[maiden geyser]

200% volume award

[gritty dust]

[hoary lion]

absolute cinema

finally made it

now I hope my windows-running pycharm works

[gritty dust]

pycharm based

[hoary lion]

rare #programming individual not trying to doxx me because I am not using neovim??

[gritty dust]

Lol

I could 👉 👈

[noble zodiac]

use whatever you want

[midnight sigil]

that looks like it got over-fitted ngl

[stark needle]

I have already doxxed you 312 years ago smh

[hoary lion]

where is the loss graph

[hoary lion]

10 generation ahead ahh

[stark needle]

Your real name is hyper.tech, not "Job application Osage"

[hoary lion]

nahh its over chat

im packing up and escaping somewhere in antartica

[opaque sigil]

[gritty dust]

Do it

[dusty niche]

gexagon

by karma ur SvelteKit and php developer

[hoary lion]

Svelte AND php??

Two things that I never thought would go along

[opaque wharf]

Not just Svelte even, SvelteKIT!

[rigid snow]

i assume laravel
or the "fullstack framework as a frontend" approach if it is sveltekit, thanks nextjs

[trim valve]

can anyone here think of a good reason why I shouldn't just buy an intel arc b580

very heavily considering it

[rigid snow]

order $250 worth of mcdonalds instead

[trim valve]

so like 3 burgers

[rigid snow]

us economy

[trim valve]

I live in the uk

[rigid snow]

british economy

[trim valve]

why does firefox limit itself to 60fps when I open it on a 60hz display

and then never increase it when it moves to a better display

[maiden geyser]

maybe you're doing it without respect

[trim valve]

I love firefox

[opaque sigil]

i don't have that

or is this about plugging in external displays and moving the window over

[green iron]

if there's anyone using cloud storage providers for longterm storage of >10TB of data with only rare retrieval, which one do you use?

[rigid snow]

idk something similar to s3 glacier

probably

[opaque sigil]

the cheapest and most reliable option would probably be s3 glacier with rclone or another frontend

yea

[trim valve]

if I have just a 60hz virtual display on, launch firefox, switch displays to my 144hz physical display then firefox stays locked at 60fps

[opaque sigil]

funky

[trim valve]

also steam really dislikes getting switched about and resizes itself to be completely full screen (overlapping the taskbar)

[opaque sigil]

i guess if they set some stuff at startup then that makes sense, though surely there's an event you can listen to when a new display gets plugged in or sth

[trim valve]

¯_(ツ)_/¯

I blame my probably very scuffed setup

its not too bad of an issue anyways, 60fps isn't a huge killer or anything

I just trolled by low framerate scrolling

[opaque sigil]

is killing the tasks and then restoring the session not an option

[trim valve]

I usually have a shitload of incognito tabs open

[opaque sigil]

ah

[trim valve]

its kinda hard to explain why I use them

[warped narwhal]

What for?

[trim valve]

literally just googling stuff

[warped narwhal]

Uh huh

That's what they all say

[rigid snow]

i do that too i thought i was schizo wtf