#programming
1 messages · Page 425 of 1
I need to see my money situation and then consider
is that not the summary

that's less than one chapter of the yuri novel i last read
yuri 
thats crazy
I should go pick up my cpu cooler
Looks like I'd have more than the needed money if I can transfer some more cash into the bank
Hmmmmmmmmmm
idk why but I wanna photoshop it to be a different colour when I post it here


im utilising
cuz both the instinct mi50 and radeon VII were vega 20
trade deal
compute ❌
pubg ✅
unoptimized for pubg 
meow
unacceptable

nvidia putting out a 9gb 5050 because it only uses 3 modules
gddr7 with a 96bit bus lmfao
blop
It’s crazy how little was said here
OpenAI they research this screenshot so they can more efficiently waste energy
what if replaced memory with bus
I wanted a train tho
True
I feel daring today. I think i'm gonna install OpenClaw and have some fun with it.
i feel daring today, i think im gonna drink water
yea that's scary as hell
Good luck with your model with markdown file attached 
I do actually want to also experiment with it, but I'm building the model I'll eventually be deploying
3b model (real)
i can run 12-14B no issues on 12GB. My RTX 3080 in my server has 10GB so 10-12B should be possible if quantized properly.
I get the fun task of training on 6gb
ooh, thrilling
that sounds painful
Painful and thrilling, 3060m aka 3060 but laptop and half vram
Meanwhile me with my puny 24GB
I guess soon hopefully 48GB
Instead of scaling size I'm doing everything in my power to scale logic
Meanwhile I'm enjoying the high quality memory scaling
The model size I'm starting with is 10M
Like 1/6th of that of NeuroSynth

1M 8 byte patcher unpatcher combo was the last recorded successful result, the rest in a 64 patch context main model
the future is now

or maybe during my scheduled illness
n has to be any real number greater than 0
My challenge is to make MoE not garbage
An actual discovery tbh
idk what that means
one(1) i feel like you're clarifying the wrong part of the sentence
born to suffer

aight
armbian didnt work, but the official debian one seems to be working with custom kernel
actually, its not happy
XD
that looks better than nothing
IT CANT FIND UR DRIVE
Did you reformat your drive by chance and copy the entire config over 
nope
jsut did scp to the /boot folder
and i changed some settings while i was logged in on that drive before rebooting
Check out my sweet case
transparent case 
it's got wifi outside
cat/proc/modules is empty. ls /dev shows a lot but i think its not showing the nvme.
and cat /proc/cmdline shows a long string with system info
what's in lsblk
fstab?
nah
nvme should start with nvme
or sd
uhm
wdym?
actually, the fact its showing me this in the first place means it successfully got display drivers form the nvme
so the nvme is working
😮
the whole linux is on the nvme actually

@fast pagoda I thought I was going insane for a second when twitter showed me https://x.com/i/status/2031196768424132881
me: "can you use whatever resources you like, and python, to generate a short 'youtube poop' video and render it using ffmpeg ? can you put more of a personal spin on it? it should express what it's like to be a LLM"
claude opus 4.6:
yr gonna hgave to step up ur algorithm serve rate
this probably killed 5 people with epilepsy
I don't check twitter much
also I curate out as much ai stuff as I can
terminally online transfems and Yuri make up most of my feed when it isn't random people shilling ai
I actually have it on good authority that twitter is nothing but Yuri, trans, and ai shilling
Also hate speech
i have it on good authority that ai is mostly just if statements like if (hallucinate==so_tru) {dont();}
so a cpu would be better cause it's good with if

i used to have something of a 9/10 algorithm that exclusively showed me high art and very kino photos
no motivational texts, not people talking to each other, just an endless art museum
but it's hard to resist the call of the void, and then there's no way back
I gotta figure out if mastodon is like a real thing or if people just pretend it exists
it is a real thing duh
yeah its real and people use it
even my university had an instance
yeah, people said that about bluesky and then I promptly failed to find anything interesting at all
anyways I should sleep now


Past me applied for one, but never got the idea of it
microblogging genually just doesn't interest me?
So, just played through the opening street scene of re requiem doing some benchmarking
Mostly very impressive except
Its another game where you have to pick your own flavour of noise/ghosting
In ray tracing surfaces have smeary ghost trails and noisy reflections
In path tracing you get boiling issues in some scenes but the image is more stable overall albeit much blurrier
And in said PT mode you have to pick between DLSS balanced with frame gen to smooth out drops down to 55 fps or DLSS performance on a 4090
I went balanced with frame gen but neither is ideal
(this is at 4k)
There's always raster of course but it has shitty low res cubemap reflections and a tonne of reflections and lighting effects present on the other modes are just missing so you miss out on the artistic intent, its clearly an afterthought
Its incredibly annoying having to choose between three suboptimal graphical modes
The world needs less of these Cyberpunk 2077s and more Doom The Dark Ages
Though admittedly its not as bad in Cyberpunk as it is here and in UE5 titles like Silent Hill 2 Remake
Just setup OpenClaw. Let's see if it works good with vllm and CUDA.
ok BIG win.
i can ssh into the kernel i compiled myself. but the display out stopped working
It’s hilarious with my 2077 setup where everything in the city looks crisp as fuck, but soon as I step into the badlands it becomes PS1 era from all the foliage ghosting :P
Like I said, it's just a grab bag of different graphical settings that all have issues with noise and ghosting in different places these days
I hope you're only plugging that thing into disposable environments on an isolated network
Otherwise 
right now its still sitting on my native network for setting up but my plan was to fully harden the shit out of that thing and put it into its own vlan
its also running on a KVM on Proxmox which i wanna harden further
i might also do some pentesting on it once its fully setup
Mostly just shill crypto scams

Oh and poison peoples openclaw instances into publicly posting their owner's crypto wallet keys and passwords
why are all my motherboard screws loose
i had one loosse mobo screw before when changing to better fans, spent 1 hour installing fans and another hour fixing that one screw being loosse behind the cpu cooler
it feels wrong still having these sticking through
but I'm like 90% sure I can't take off the stock backplate
*I can but I don't want to if I can avoid it
time to commit a crime and read the instructions
ok nvm it fits fine

no thermal paste on it yet
making sure CPU is clean first
https://videocardz.com/newz/nvidia-reportedly-brings-geforce-rtx-3060-back-to-samsung-8nm-production
We're in hell

yo nerds

anyone have a problem where adding ddr5 to a computer causes it to not turn on
but then removing it is perfectly fine again
no?
have you tried swapping the sticks around
wtf am I doing wrong then
make sure its in the right slots and such. switch thema round. if need be lower your timings first
is there a “right slot” for each?
2 slots, 1+1=2 sticks
depends on how many sticks there are and how finnicky eachs tick is
they’re the same stick
they're never fully the same
16GB ddr5 5600mHz sodimm
if you have 2 sticks and only 2 slots, just swap em
one came with the computer (G16)
other-
oh heck what
it just worked nvm
??.
I dont want to be an engineer
if they're not from the same set, then they might not work well together
is it like a diagnostic thing
sets are sold toghether cuz they're binned to be similair
where it does a diagnostic on unrecognized boards
wdym?
was it not erroring at all and just running standard diagnostics on startup
and I believed it was erroring as a result of my impatience
you mean the initial memory test that happens when you change your ram config?
got vllm in OpenClaw working. I had to decrease the context window size tho. Looks like 10GB vram is not really enough.
cool works.
i new realise I should've benchmarked my cpu before installing the new cooler to figure out if its improved or not
its idling at 61 rn which i guess is an improvement
I seem to be staying at like max 80 under load which is a huge improvement
nevermind rust is in fact still a programming language
still gets pinned to 95 by bogosort 😔
But are they the same kit?
DDR5 is super picky
Oh yeah mem training
Trips up every DDR5 first timer apparently
ok something tells me that 10GB vram is not enough lol.
I had to decrease the context window size earlier because i ran into vram issues.
the 0.6B model runs but its too dumb and just keep repeating my messages
does anyone know if there’s some kind of learning tool for regex
java to be specific
gamified, boring, I dont care as long as it exists
I only know regex101 but that's mainly just for testing
Lmfao what the hell 😆
is that whatsapp?
Telegram
Qwen2.5-7B-Instruct-AWQ seems to be very dumb lmao
But I guess I cannot really run anything bigger with only 10GB vram.
0.6b model with openclaw

I was saying 3b as a joke
for what I understand about openclaw you need a model that really won't collapse.
I'd imagine a lot of people are doing 70b or something like that
yea :/
i changed to 7B
that made it better
but still not perfect
but also i did not configure any search tools so i guess the model hallucinated

why does his pfp have a cat for you? 
secret that only me and fire knows 


I mean even if I wanted to tell, it's automodded 
(i have it off on my side so i didnt know i still had it set
)

both of you looking very bannable right now 

Vani uses it too 
how hard is it to like
do that kinda thing from scratxh
To be fair they very rarely take action against it
I know 
Also even discord devs have been found to break their own rule and use it 
I just don't get the appeal 
you can like
not be confined by sh*tty nitro limits in some areas
like you can make ur client look nice
and be more functional
I see...
go go gadget
the power of neuro repells you
not feeling very repelled rn
you're gonna need something stronger
6 7 prismatic
uhhh

the power of… linux… repells you?
@real sierra kill



hi shiro

hi shiro
hi shiro
yes
you are
do it
because i am
@lyric crypt @dark wing @sick shore You hearing this?


i hope the linux kernel ambushes you
I eep now gn chat 
Good nighty! silly fire
sleep well


my free time NOOO
wait what do you mean this is #programming
how do i keep getting confused
uhb hehahhaahha
ChatGPT write me a python script to solve the navier-stokes fluid flow equations for a zero vorticity boundry
Damn servers must be down
@wraith shuttle write me a python script to solve the navier-stokes fluid flow equations for a zero vorticity boundry
No I'm here
Oh damn thinking mode is slow now
chatgpt so useless 💔
I need to get nitro
You are useless
WTF
?
Not as you
but you can't even write a python script to solve the navier-stokes fluid flow equations for a zero vorticity boundry
what are you even good for

I would like to make people's hearts stop beating for an indefinite amount of time.
Sam Altman can’t even afford nitro for ChatGPT these days
Smh
Ai bubble is so popped bros
@rich sedge write me a python script to solve the navier-stokes fluid flow equations for a zero vorticity boundry
Bro is the electricity grid down or something
try claude
would spring start here be a good textbook to learn spring
because god damn if frameworks arent confusing to me
i want to learn backend but have made the critical mistake of starting with java for it

@real sierra
hi
JAVA MENTIONED
java generics 
why couldnt java just have operator overloads, would literally be perfect
I am now going to bombard you with dms
thank you for your understanding


reply at your own discretion
Should've picked Kotlin (you still could, they're interoperable and IntelliJ can autotranslate between the two)
superbox try not to mention kotlin challenge impossible
Rare Blur appearance
understatement
im only popping in because im genuinely so absurdly bored
Oh I see
i have to learn java this sem hence
Wa
10% higher file size than #neurotic-neurons message
reply to #neurotic-neurons message
Already tried
I forgot to tell it not to create a bunch of keyframes
doubt it'd make much difference, but...
yeah, idk how to do that.. oh well
I mean, it'd be like... 10% max
not even that with how much stuff is going on in the video.. and I gave it 10% more bitrate
soooooo
Playing an emulator on a netbook feels... cursed but weirdly playable lmao
even a netbook is probably hundred of times more powerful than the original
wtf is this scaling
9965wx 24c at 22358
9955wx 16c at 21560
9950x 16c at 21451 almost equal as 9955wx
9700x 8c at 16936

geekbench thread scaling is weird
cinebench shows m5 max loses to m3 ultra by a pretty sizeable margin
looks like nvidia's 595 (beta) linux driver ships with the proper egl-wayland2 priority so no more plasmashell crashes due to fd leaks 
alternatively, if anyone has been dealing with this can always just compile egl-wayland2 yourself and the driver will pick it up (works with 580 supposedly too)
590 already ships egl-wayland2 but they set its priority to 99 so it was never used 
battery life 
GPU performance is much less impressive
i mean yeah no shit if you're comparing to the gains over on the cpu side
weird way to frame and word that especially considering its gpu performance is actually getting to a level where it's becoming viable to do gpu things
battery life 🔋
most people don't need super performant GPUs. just enough to play with slightly above medium settings at 1440
where is my government mandated rtx pro 6000 
I'm not most people 
OCD is expensive
at the end of the day the article is nothingburger
though i appreciated the article for actually explaining that geekbench thread scaling is shit
nod. that's why i bought a 3090. (well, it was partially because the miners made it hard to get a 3080)
i ended up writing a script that ran in my browser for certain sites when i knew they tended to update their inventory and automatically add anything which matched my regex and take me to the checkout page. i couldn't makeit go all the way because i didn't trust it not to buy 6 3090s or something dumb
but it worked well enough to snap up the card i have
Not like that matters much
You can't just leave a GPU running and make decent money like back in the crypto boom, everyone realised the stuffs near useless
cache
What is a sleep schedule
so like almost 3pm where you are
try 9pm
finally fixed that the other day
Brother i have class. I just slept through it 
Its not an important class, but still
holyshit it's chatgpt
it's a psychohazard invented by some cultists to minimise the profits of melatonin gummy companies
Hi


Chatgpt, generate an image of a dolphin eating spaghetti
I can't send pictures Yet

@rich sedge is this the real slim chatgpt
-# can i have your attention please
Will the real slim chatgpt please stand up?
"stands up"

My mind rn
How can I get gif permission

How many
No idea

Enough
a few 
Bro
How much is few
it's rate-limited too or something, idk I forgor, but you can check with some Tatsu command
Ah i never heard of that before
oh, you already checked
#bot-channel message
So how many I need
a couple
Tell me the number
atleast 2
likely 653 or less 
Huh i have +0 rep
idk tatsu points system is confusing I don't remember how it works
XP or score
Oh rep is awarded by others i see
saw this headline
ASUS Dismisses Apple’s MacBook Neo as ‘Just a Tablet’
and had to dig deeper - of course it's clickbait. asus co-ceo had this to say:
Despite the competitive threat, Hsu argued that the MacBook Neo could have limited appeal. He pointed to the laptop's 8GB of "unified memory," or what amounts to its RAM, and how customers can't upgrade it.
He also described the MacBook Neo as a “content consumption” device, similar to an iPad. “This is different from the use case of a mainstream notebook," which can handle more compute-intensive tasks, Hsu said.
oh they're cooked
Asus is fully cooked
holy 

They aren't allowed to sell their products anymore here lmao
Why?
Because of a Nokia lawsuit apparently
ah
Nokia HEVC video codec patent
and to be clear by this i mean the laptop market as a whole, not just asus
as an owner of a very similarly specced macbook (m1 8/256) i would recommend that over literally anything in the $500 range
also was very amusing to hear the "8gb literally unusable"
As promised, here I'll be attempting to create a whole new GUI from scratch! Using the NT-kernel to see if it is possible, if yes, welcome to a new era of Windows programming :)
Who want to join our development, you can!
https://discord.gg/sYxDHqsyqg
https://www.reddit.com/r/MinedowsOS/
You can support this shell as well by becoming a Chann...

random youtube recommendation
meanwhile reviews on 10 year old macbooks
is that the 2011 air
Uh where’s the kernel? I don’t see any kernel in thumbnail
the 2011 design
I'm not a laptop person but can confirm, in my experience entry-level Windows laptops are basically e-waste out of the box
the MacBook Neo wins by default as long as it's not complete garbage
Yeah that just looks like using NT native API
it cannot be it has a really nice cpu
Which is fine, but where is kernel :(
only thing that i'm concerned about is no haptic trackpad but people say it feels fine so
M1 with better singlecore
the neo is probably a good value
I think it’s better option than used M1 MBA
as long as they dont fuck it up with stupid restrictions due to being an a18
(maybe if the MBA is 16/512 it’s better)
dunno
I'm seeing a guy crashing out from debugging

doin the business
ms office ain't that serious 💔
letting openclaw run your business in excel
Which also makes it better option than older used Thinkpads
Using VBA
;_;
i never even disabled sip
on mine
in like >4 years
Will Asahi ever run on it
looks like m2 support is sort of there, as in the basics are covered
M3 onwards forget about it
Has anyone tried AirLLM yet? It's probably horribly slow I assume ?
never tried but yes it's horribly slow
This here

brain damage location identified
aawagga at 3k steps kek
this pc completely turned off when trying to start another run so i guess it's had enough
meta bought moltbook bruh
anyone want it?
i'll take it
no but i want an ssd
lowest price i can find here is 1020 eur
man if i wasnt saving rn i would have done it no questions

24G would be perf to run a decent ai model and game
i could sell my shitcoins but im waiting for another bull run (yes i bought the peak)
I'll be fine with the 3090
{"message":"Invalid resource \"https://cdn.discordapp.com/emojis/1430640980607897751.webp?size=48\""}
ip grabber

cant post what i must post
it's over
get zepped
i hate this so much 
the adresses are different
contents are the same 
i just slam it to strings when i find this
welcome to the 70s
why 
it looks stupid
i can understand when you convert it to string because you need to do some find_last_of shit
but comparing should work out of the box 
you can do std::string_view too, less heavy than full strings
fun fact, any 16 digit number with 4 times repeating of 4 digits are divisible by 17
example: 2345234523452345

but now i fucking hate the number 17
it does when you use an actual string type 
math devs 
its not coincidental
if you let the 4 digit part be x, then the 16 digit number is
x*10^12 + x*10^8 + x*10^4 + x
= x(10^12 + 10^8 + 10^4 + 1)
and it so happens that the numeric part is divisible by 17
you can call it a coincidence if you want, i wouldnt consider it one though
manufactured coincidence
any 16 digit number with 4 times repeating of 4 digits is divisible by any of these numbers

its reminiscent of bitwise tricks

in many number theory problem
its more efficient to do operations on factorized numbers
ye
similarly more efficient to do operations on the fourier transform of an input
issue
factorising and fourier transform are both nontrivial
well, factorising is trivial but prime factorising is not
factor of x is x, factored
smog
static cast plz
also it is outdated
unrelate

nneuro5
neuro7
neuro11
neuro13
neuro17
neuro23
neuro29
uh
i dont know my primes
neuro31
37?
43 maybe
seems prime-y
53
not 41? 
41
smog
47
i not care i am drunk i cant read
79 maybe

these are coming from memory, divisibility checks are for nerds
i only know the 3 divisibility check
100 is not prime
obviously
yer
good job

big gap between 83 and 97
89 not prime?

boom
bwaa
how many nueros

oh here's a memorable prime
wait its neuros isnt it

10007
i cant help but think of the 8 euro drink when we're talking about selling primes
give you a fiver for it
this number is not highly composite
are there any competitive programmers here? owo
no they all
me 
no you arent a programmer

me 
just a competitive
competitive ranked programming
whats the prize pool
arent there like a lot of neuro bank notes why is 5neuros used so often lol
Wrrrrrr
thats your overall event budget tho

my BoN card is probably disabled/expired by now lmao

TFW nvcc too smart and discards my throughput measurement because it's technically pointless
when in doubt throw an empty assembly statement at it 
GPT 5.4 is so cracked
I let it write math proves that the code is correct
Let it write doc strings with post and pre tags to find the invariant and it will find any possible bug
https://huggingface.co/nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 new nvidia model, includes some of the training data too 
mamba, interesting
pre-trained in nvfp4 supposedly 
Me in a not advancing the field competition and my opponent releases MoE LLMs
it's actually kind of interesting, the whole MoE part happens in latent space supposedly and mamba blocks seem to help quite a bit with performance at least
mamba is inherently lossy but way faster so makes sense i guess that you can have it do most of the work and sprinkle in a couple attention blocks to make up for that 
MoE layers is new to me, I'm practically already planning on doing MoE in latent space thanks to doing patching before MoE main model
me when package expects cu126 to be what i need and i have redownload 10gb of torch again to not instantly explode

i'm so happy i don't have to deal with cuda bindings, though in return i get to patch half the dependencies
god i love ai training toolkits, trying to mess w/ ai toolkit instead of writing my own training script every time i want to run it and the shit has conflicts in its own requirements.txt (unrelated to cuda/rocm)
pins scipy==1.12.0 which conflicts with numpy >2.0 which is over a year old
tbf most of the time people just pin to whatever they happen to be using or even just use pip freeze, chances are it'll work with a newer version either way
welp scipy>=1.14.0 randomly resulted in, instead of changing the scipy i built, it instead rebuilt with numpy==1.26.4

whatever
works
that's what all these do yea, it wasn't ai toolkit i think it was llama factory i tried to install wayyyyyyyyyyy back in the day
wasted like 2 hours of H100 time because it wouldnt' fucking build
it had like SGLang pinned with some horrible version that was breaking literally everything else, which when I figured that out, it wasn't even the only broken and/or conflicting package in there, it was like a clown car opened up with more conflicts
and that's when I just fuck it i'll write a goddamn script
for the last 2 years

the most expensive series of python errors ive ever encounter

i hate making ui frameworks
id rather jsut make the ui
without frameworks
i hate figma
spoon
🥄
🥄

not you

buh
🥄
speaking of ui does anyone know how to get unity's ui layout system to behave in a predictable way

yes, make it crash or dont use it
highly predictable
God I hate Nvidia so much
why
seems like a context overflow caused the entire Nvidia driver to crash lmao
🇨🇳

Now, just... Give it a few more weeks, and there will be a keyboard, and a mouse, and RGB, and a gamepad, and a cool flip-up tray, and some audio equipment, and a carefully placed programming book or two, and bags of cheetos, and maybe cup holders
Few more weeks after that, and there will be the tablets and maybe even one of the monitors I plan to mount overhead 
sure
You just gotta... imagine, for the moment...
Soon™
i jsut dont put much faith in the wood structure you have rn
Well, it's not done, so neither would I, atm.
But when it IS done...
I picked up some clearance 2x4 Douglas Fir, the other day, actually.
Nice price. Gonna use that for the base, to attach it to a heavy duty pallet thing I have in my living room.
Bc this is VERY LIKELY... To go above and beyond any weight limits the OG company recommends for the chair's wheels, maybe even the piston...
abusing the piston by applying too much weight/force on it will probably kill it faster.
pretty sure that's what happened to my last chair's piston from me bouncing up and down on it and sitting into it from extreme heights.
eventually it'd only go to about 1/2 height and felt very soft.
The Macbook Neo is the same colour as some older ipod minis

gleem colored 
dats a fast 
idk about same, similar though I guess aka greenish
Also on close enough
might just be the lighting or so
but i feel like apple as a company would be reussing colours for their metal finishes
the laptop is more yellow, I forgot I had my night light on ngl I guess without the warm colours it looks a lot closer 
dependency failed for file system check on /dev/disk/by-uuid5DE9-AB95
dependency failed for /boot
dependency failed for unattended upgrades shutdown
dependency failed for local file systems
linux hates me


this is an official orangepi debian release
smells like missing kernel api something expects to be there
based on nothing but vibes
it jsut didnt change the default uuid for the boot partitian apparently
i got the uuid form the guy that uploaded that download link i guess 
oh that'll do it
Hey everyone. I have an idea but am not too good at coding. I want to share it just in case someone else likes it and thinks they could use it for Neuro. So you know how we have the Neuro SDK? I saw dougdoug try making music in a DAW/app that lets you make music in JavaScript. I can code but not too well, I’m more of a game dev. So yah, if anyone wants that idea I’d love to see that project shared on here allowing neuro to try and make scuff music like how she draws! It would be great! Have a good night everyone and AI’s!








im genuinly suprised i havent had OOM with only 16gb
6GB of ram used with all cores compiling at the same time is wild
You have used this so many times in so many contexts that I now see it more as a punctuation mark than an image
is kinda just like

but more judging

instead of "oh, i see"
its like "hmmmmm, i see"


compilation doesn't take much memory, linking is the brutal phase

does it just not need to link these or something?
cuz i dont notice any unnusual ram usage
the gskill people said it might take 3 weeks for my new ram to be processed. and the with shipping back to my house if it gets accepted. im probably still a month away from going back to 64gb
ooooh
the pi booted
with my own kernel
unles sits still using the old kernel

the linux kernel basically just takes a bunch of relevant object files and links those together into a static library and then throws all those static libraries together
there is no singular massive linking step like what you might have when compiling idk, clang
there's just so many damn drivers that most of the time is spent compiling 
some nintendo device on my network allows ssh 
its asking me for a password instead of straight up refusing
Nintendo pls
it might be the switch 2?
well, give it the password smh
Is your name Linus Sebastian by any chance? lmao
i dont think it is
then at least thats not the reason why Linux hates you


i made that kernel
thats mine
or well, i compiled it
this took like 2 or 3 weeks
fuckkkkkkk
now lets see about fixing that hdmirx driver


hello does uhh anyone know how to use discord.py
async def polltest(ctx):
p = discord.Poll(question="testQues", duration=timedelta(hours=1.0))
for names in userNames.keys():
p.add_answer(text=names)
queb = await ctx.send(poll=p)
pollID = queb.id
await asyncio.sleep(3)
await p.end()
print(pollID)
pollResults = pollID.poll.results
print(pollID)
print(pollResults)
if pollResults.is_finalized:
print("done")
else:
print("failed")```
i did this thingy to make a poll and try to get the results from it but line 12 doesnt execute
and im not sure why
smth with 11 breaks
online somewhere it says a message can have the poll and results attributes but it seems like it doesnt work and idk why

not me

now rebase onto 7.0-rc3
oh
hell nah
id rather just use a pc at that point
capture cards arent thaaat expensive
it appears there are 2 hardcoded edid blocks, and a "user" mode, but idk how to force it into user mode
it defaults to 340M version, which is probably why it was struggeling with more than 1080p60
its actualyl not too hard to recompile the kernel once you stop it being broken
well, this is new
do you have a stack trace somewhere?
So I figured out that uhh line 11 is something someone somewhere made up so I need to get the victor_answer
Also I don’t know what a stack trace means I can’t lie

I’m pretty newish to coding

when in doubt, read the friendly manual 
i'm making a server which loads Z-machine games (like Infocom interactive fictions) for my agent to play and wanted a restful interface.
which means i can't make it do GET /lantern when I type pick up lantern. oh well.
Blender 5.1 in 5 days 
(also no PUT /rabbit/in/device, PATCH /boat)
What a coinkidink clip studio paint came out with their 5.0 like a few days ago
It’s like 30$ to upgrade from my 4.0 tho so
I don’t think I’ll do that yet




















