programming

[obsidian mantle]

(todo: define performance)

[rigid snow]

ok right you kinda can but you can detect that even easier

profile acc <75%

or not profile acc

[obsidian mantle]

yeah so if you play a map but click nothing its 0% acc

[sage crag]

we dont have play time for the current month

only total play time

[obsidian mantle]

you can store last month playtime

[rigid snow]

snapshot playtime today, compare to next month

[sage crag]

waste of time

[rigid snow]

what

[keen hatch]

Yes, I tested it. Overall, k2-thinking is better than the original version, but it has all the same problems as the original – a tendency to oversimplify and compress everything, and the model doesn't follow instructions well (the model partially ignores prompt instructions, and a couple of times it executed all the secondary tasks while ignoring the main one). But overall, it works better. In my case, after adjusting the prompt, the model technically completes the task with a 40% probability (the validator found no errors. and the model spends an average of 12,000 tokens per task). DeepSeek V3.1 Terminus does this with an 80% probability (because it follows instructions better and doesn't strive to optimize everything) in non-thinking mode. Perhaps if I experiment with the prompt more, I can improve the result for k2-thinking.

BTW: DeepSeek R1, Qwen3 Coder 480B and GLM 4.5 can't complete the task.

[sage crag]

playcount is more important than playtime because statistically people with high playcount are retrying maps more

[obsidian mantle]

true true

[sage crag]

people who are retrying maps more are getting more ranks a day on average

[rigid snow]

that’s not a food thing

[sage crag]

so they increase the team ranking

[sage crag]

osu mechanics

[rigid snow]

explod retry spammers

[obsidian mantle]

so
(rank_change)/playcount

and something else

[sage crag]

there is rank history data in the userdata but only for the last 90 days

[obsidian mantle]

so 3 months

pretty much enough

[rigid snow]

why isn’t that enough

once again snapshot snapshot snapshot

[sage crag]

i mean i can very easily predict future ranks, but ranks arent what i want to measure

i want to measure performance points

[rigid snow]

ranks reflect that

[sage crag]

ranking = sum((
    data["supporter"] * 1e8,
    math.log1p(predict_playcount(data["monthly_playcounts"])),
    math.log1p(data["pp"]),
))

the ranking system right now is very naive

[obsidian mantle]

whats log1p

[sage crag]

not consistently

[sage crag]

log 1 plus

log(x + 1)

[rigid snow]

blahblahblah

[obsidian mantle]

why do you want to predict

[sage crag]

because the goal is growth

[rigid snow]

i want to shrink

[sage crag]

if predict good => growth

{
  "username": "Flyxuated",
  "supporter": true,
  "pp": 601.791,
  "monthly_playcounts": [
    415,
    275
  ]
}

here's the data i actually locally store when im gathering userdata

i use it in kill.py, but i can add more arbitrarily

[obsidian mantle]

is it true that rank directly correlates with pp

more rank more pp

[rigid snow]

this true

[obsidian mantle]

ok then you can use rank as pp substitute

[rigid snow]

that’s what i’m saying

[sage crag]

you cant though

rank is more volatile than pp

its not as useful statistics-wise

pp is stable rank fluctuates

[rigid snow]

yeah sure

[sage crag]

[obsidian mantle]

its stable enough month-wise

[sage crag]

[rigid snow]

when pp update it does

[sage crag]

not just when pp update

[rigid snow]

[sage crag]

if people overtake you, your pp remain same

rank decay happen but pp remain constant

[obsidian mantle]

doesnt matter

dont play - rank goes down

[rigid snow]

rank decay is fast enough to matter

[obsidian mantle]

then you will filter people who give maximum rank decay

which is the same as comparing pp change

[sage crag]

[obsidian mantle]

it doesnt matter how you measure it, if you measure it the same way for everyone and it correlates with pp growth

[sage crag]

i feel like i just explained why it doesnt though

[obsidian mantle]

pp can remain the same, rank will go down
pp can go up, rank will go up
pp can go up, rank will stay the same

i kind of lack of understanding what pp is

can pp go down

[rigid snow]

yes but in very rare instances

is that even still true i dont play the game anymore

can you have bigger score but less pp with csr

you should but harder right

[obsidian mantle]

so if pp cant go down then rank is even better

gives more information on performance than pp

[olive sable]

idk what we're talking baout

oh today is the 3d debut

i forgot

holy fuck its already saterday

time is a blur

[olive sable]

lmao she's hovering above the stage

[dry crest]

How does it work? Like how does she controll the body?

In the stream

[patent shard]

Is it true that if a gel battery gets too low, it destroys itself
What the fuck kind of design is that
You pay more to destroy your battery, sounds legit

[real salmon]

"bonnie does not abide by the laws of space time or physics"

[amber fractal]

I missed the debut

[olive sable]

according to vedal there are no premade poses

[amber fractal]

I'm guessing I'll have to expand the FAQ now

[olive sable]

i have 48K channel points apparently

i go to stream so infrequently thast everytime i see it im suprised

[amber fractal]

gamba

[olive sable]

its from the subathons i think

[nocturne olive]

Welp time to add 3D control and just put "No" in premade and "Figure it out" in DIY

[amber fractal]

Cloning the N row makes sense

As there is sorta an entire industry around this

[nocturne olive]

Sil

[amber fractal]

Before I forget

While I'm updating, I want to take LLM recommendations to place it in.
If anyone is recommending anything, here are the weight classes I'm considering via vram requirements, 3gb(in case someone is doing CPU for some reason) 6gb, 12gb, 24gb, 48gb.
I'm not going to be strict besides it can't be useless for any regard.

Shame that I don't really use LLMs to fill in this section myself

https://github.com/igglyn/neuroprogrammingfaq/ I sorta own this particular repo.

just in case anyone was wondering

[olive sable]

oops

[amber fractal]

sam typo

[olive sable]

i pressed the wrong emote

the skull emoji is right next to neurovbounce in my favourites lmao

[amber fractal]

neurobounce isn't even in my favorites

[amber fractal]

I need a category called "what is actually used by neuro" so I can copy-paste WE DON'T KNOW in all of them

[nocturne olive]

So silly

[mighty thorn]

Absolute bull

[olive sable]

i mean

it did look pretty not-scripted

[nocturne olive]

Also we still need NeuroSynth somewhere on there

[mighty thorn]

If it was actually synthetic, it’d be used in her 2d model

[olive sable]

nah

the 2D and 3D models work very different

[mighty thorn]

(I’d like it to be known that literally the only part of stream I saw was Neuro falling down the stairs)

[rough bloom]

we do kinda know the TTS (for Neuro only)

[amber fractal]

Yeah, the actual chat faq has been sidelined for awhile due to needing the AI section

[amber fractal]

True

[mighty thorn]

If only someone here had recent experience with a 2d equivalent like audio2face

[olive sable]

well so basicly, she broke her neck as lot when looking up and downm, and rotatingher neck 180 degrees

[mighty thorn]

Ahh but alas

[olive sable]

also just randomly started floating

[mighty thorn]

‘Twas not to be

[amber fractal]

Ascend

[nocturne olive]

Silly

[olive sable]

you cant directly use a system designed for 3D movement in 2D

[mighty thorn]

Sure you can

Just flatten the 3d access for bone positioning

It’s known that Vedal uses a custom C based live2d engine

Couldn’t be that hard

[amber fractal]

[mighty thorn]

If he can move each individual sprite of the model

[rough bloom]

[olive sable]

the problem is that flattening it would mean scaling things weirdly

for example if the arms point directly at the camera

thats one of the problems i see at least

idk what vedal is doing behind the scenes

[amber fractal]

For the few things I know, live2d is pretty much uv wrapping a 2d object in terms of how it is animated

[rough bloom]

tbf it doesn't even matter if it's possible or not (since it probably is, you'd "just" have to train whatever model is currently used for 3D with the Live2D model and parameters)
2D and 3D have totally different requirements, being able to roam around and do stuff is needed for 3D but not for Live2D

[olive sable]

i think he could use a similar system if he wanted to, but not a direct port

and he hasnt

probably keeping the 2D system a separate black box

[amber fractal]

Hear me out, we have neur in a desk chair and reality breaks into 3d as she steps out. Probably just disabling a lot of calls to legs to even consider making it work so it is useless.

[olive sable]

we cant really know for sure unless we get a vedal q&a somehow, but he hates #programming

[mighty thorn]

Me when I’m in an indie project public transparency competition and my opponent is vedal987 (I forgot to show up and still won)

[olive sable]

neuro is very much closed source yep

[cosmic sphinx]

idk if its 100% true, cause she did several cart wheels in a row practically flawlessly
BUT, I wouldnt say they were 100% identical either, so maybe if its a neural network that does the actions, it does have the data for that somewhere in training

the quality of this model is actually good tho

[mighty thorn]

Bro literally just released the among us plugin and decided he has repaid his intellectual debt to our species

[obsidian mantle]

Naaah it was 100% premade with direction variables maybe, and different starting positions it gave that variation

[olive sable]

im pretty sure i heard him say there were no premade poses. either im going senile, vedal is lying, or neuro is built different

[mighty thorn]

It is very much possible that he borrowed / copied filians equipment

She has really advanced stuff

[young plover]

vedal lying? never

[obsidian mantle]

Same for spinning

[olive sable]

[cosmic sphinx]

its too smooth not to be premade tbf

[obsidian mantle]

She has a set of commands to control it probably

[olive sable]

with sincere regret, and a promise of improvement, your name

[obsidian mantle]

With ability to move it freely sometimes

i cant figure out why he just didnt detatch the camera when they played hide and seek

[rough bloom]

that seems like what happened, yeah
it seems like she was giving natural language instructions to another model that did the movement, so there's a good chance that the instructions are incomprehensible most of the time, except for stuff like cartwheels
easily describable actions work well, complex ones not so much

but just speculation

[olive sable]

maybe its just not set up that way

[obsidian mantle]

Discord is tweaking, it says the message was sent and then it teleports to next message i post

Yeah probably not set up like that...

Anyway it looked better than i expected

Also walking is premade too clearly

[olive sable]

https://youtu.be/R0b7wqR_FsQ?si=4_mUxrdGrGHvwVlK&t=292

[obsidian mantle]

Swimming movement was smooth but i doubt its premade. Looked surprisingly good

[olive sable]

im not senile, he said its not prerecorded

tutel would never lie, surely

[nocturne olive]

I have gotten Wispers to poke a wizard, surely that means the next NS cover will get an inst soon

[woven quail]

i'm sure that it's trained

[olive sable]

ye, that's the most likely reason it works smoothly i think

[obsidian mantle]

trained? So there is a separate ai level who controls the model?

That would be sick

[olive sable]

i think something like that

[rough bloom]

half of the training data is cartwheels

[woven quail]

that should be the case here

[obsidian mantle]

So it can actually learn

[cosmic sphinx]

for most of the arm/torso/head movements, its a neural network, which likely responds with actions in timing to what she's saying
like, I noticed it especially during the songs, that she nailed some movements and jumps in exact same moment as she spoke a lyric, or went with a beat

walking, picking up/dropping stuff, which happened usually right after a particular command, is maybe done by the same model, but it requires a precise instruction from neuro (ex. walk to screen). ofc its more complex than that, just in general words how I think it works. same goes for spinning, cart wheel, etc. I dont think those are actually prerecorded, but are just apart of the training dataset, which is why she's so good at it

[cosmic sphinx]

obviously, a LLM cannot directly control a 3D body flawlessly, you need an 'interface' model

[rough bloom]

that's quite obvious I think
there is no way in hell that the LLM is directly controlling every movement in real-time with even remotely enough accuracy

[mighty thorn]

Not necessarily

I mean

[cosmic sphinx]

where has been the case otherwise

[mighty thorn]

No way he is actually doing it this way,

But image and audio output modality exist

[cosmic sphinx]

image and audio yes

for robotics, you can technically predefine concrete instructions with code functions, and just let LLM call them when needed

but for complex movements as here, you'd need a separate model

[mighty thorn]

Well no

There’s nothing stopping you from attempting to have an output modality of 3d tracking data

[cosmic sphinx]

maybe

[mighty thorn]

Just money time data and practicality

But you know

[olive sable]

me when my llm output is:
"text123456789 lorem ipsum, Larm: 15.1, 48.9, 17.1, 98.5, Larm: 87.2, 64.0, 48.5, 71.2, Lleg: ................"

[mighty thorn]

Minor stuff

[cosmic sphinx]

its definitely possible

[obsidian mantle]

He should ask filian to record some movement tracking data

[cosmic sphinx]

but it'd probably be more efficient to use a specialized model rather than finetuning a general LLM

[mighty thorn]

I may or may not be training such while passing the time in this chat

I can neither confirm nor deny

[cosmic sphinx]

[rough bloom]

idk how it works during the singing parts, but yeah, everything else seems to be instructions/tasks that are handled by another model
I think in general you could quite easily notice when it switches (regardless of how scuffed it is)

[mighty thorn]

But if I was, it’d only be via adding an ARKit output head to Qwen2.5-Omni, not fully 3d stuff yet

[olive sable]

how many bones does neuro have? i imagine at least 2 per arm, 2 per leg, 1 per foot, 2 in her body?
so at the very minimum 12, probably a lot more

[mighty thorn]

But I may or may not be so we’ll never know

[woven quail]

I'm actually glad that vedal made her able to move when she speak

[obsidian mantle]

You can probably count it based on her movement

[olive sable]

for the hands, i would geuss hes not having the model interface with each finger. i would geuss its 1 input controlling the open or close

i don't have the time to hyper-analyze the vod tho

[cosmic sphinx]

could train on a dataset, where words/sentences/phrases correspond to a body movement

and it works for both normal actions and singing

if you include both into dataset

[olive sable]

ye

[supple gulch]

I Just can't comprehend How Tutel goes from 0 -> Screen Recognition -> +Large Language Model + 2D Model -> + TTS +STT -> +model that shows emotion and gets Controlled +emotion in Speach + some how some times game suport -> +3D-Model ; Is there any confirmed information on how he does that (expect the game integration on Github), or how many Ai 's he needs for this many diffrent actions?

[woven quail]

well, for sure he's not doing this alone

[olive sable]

he has been working on neuro for years by now, so its just dedication

[woven quail]

ye, he once said that neuro is his life project

[olive sable]

there hasn't been much confirmation on a lot of these, so

[woven quail]

if not his dedication, who know where neuro would be today

[supple gulch]

not at this point, she is now, for certain

[mighty thorn]

@old pendant grok is this true?

[olive sable]

also wdym with 0 -> ...? like how she was made?

cuz it was built piece by piece ofcourse

[mighty thorn]

If I had a full time job’s worth of passive income and 2 years I think I could match 2024 Neuro

[woven quail]

I think what he meant there is that vedal knew nothing about AI at the start

[velvet crest]

You have unlocked new role

[olive sable]

im currently matching 2010 neuro levels of ai tech

[olive sable]

ah

[mighty thorn]

Hell depending on whether you interpret “match” to be closer to “duplicate” or “equate” for the things we have no idea about, could be a few months

[nocturne olive]

Meanwhile me casually basically matching 2025 Neuro singing tech

[olive sable]

superbox stop glazing your neurosynth 💀

[nocturne olive]

Sil

But it's peak thou

[mighty thorn]

Did you guys hear that Neurosynth was voted best CRISPR based gene therapy clinical trial of the decade?

[olive sable]

nice

we love cripsr cas

[dry crest]

I dont think my AI knows the law:

You: do drugs
Bot: let ’ s do it !

[nocturne olive]

Wa

[mighty thorn]

[nocturne olive]

#1 NeuroSynth fan spotted

[olive sable]

you

literaly you

[olive sable]

i think that was a reaction to the message from jullie lmao

[rough bloom]

real

[mighty thorn]

ITS MY NAME

IDIOT

[nocturne olive]

Look at the display name

[mighty thorn]

GOD

Belgians smh

[olive sable]

oh

i didnt even notice lmao

i dont look at names

[mighty thorn]

This guy

[olive sable]

i look at pfp

[sage crag]

https://tenor.com/view/uma-musume-uma-im-no-ears-im-all-ears-horse-gif-4308252428337868305

[rough bloom]

[olive sable]

hi konii

[nocturne olive]

Sadly can't make that NeuroSynth yet

[olive sable]

make a neuroynth logo then

[sage crag]

https://tenor.com/view/yume-nikki-madotsuki-gobou-poniko-monoe-gif-14969244138185379825

[nocturne olive]

Oh you wouldn't know what's going on behind the scenes

[olive sable]

i indeed dont know

[nocturne olive]

We have a lot more than just a logo coming up

[olive sable]

cool i guess

[mighty thorn]

Can you just render the weights of Neurosynth into an rgb image so I can use that

[nocturne olive]

No that would be leaking

[olive sable]

im not really up to date on whatever is happening with neurosynth

[mighty thorn]

Compression

It wouldn’t tho

Cause compression

[sage crag]

<-

[supple gulch]

Yes (i don't know if he knew a littel bit, i only heard that he is self taught) and even if he knew, he would needed not only that, he needed training data and the know how to interact with "Live 2D" i think and other programms....

[olive sable]

[nocturne olive]

Latest updates:

• Tuning 2.0
• Japanese and Finnish singing now possible
• New cover coming soon once Wispers' wizard gets me an instrumental
• More assets are in the works

[mighty thorn]

I have suspicion with the often repeated and, to my knowledge, sourceless claim that Neuro’s training data was “ethically acquired”

[opaque sigil]

leaked logo

[mighty thorn]

[olive sable]

from what i rememer he tried out gpt 3 and then went form there learning ai and stuff

[nocturne olive]

WayTooSilly

[woven quail]

ethically indeed xd

[olive sable]

learnign from 0 is pretty much how learning works tho

if you're really dedicated you'll figure it out

[supple gulch]

yes, but that was meant by 0 too

[mighty thorn]

Dw tho Vedal, I have very little concern for ethics

[young plover]

that's obviously not true
you cannot train an LLM from scratch on twitch chat lmao

[opaque sigil]

you can

[mighty thorn]

funny you say that actually

[opaque sigil]

just ignore the quality

[mighty thorn]

Also

It’s not from scratch

[zinc karma]

I wonder if Neuro or another llm is coding to control the joints. Something like this?
https://www.nature.com/articles/s42256-025-01005-x

[mighty thorn]

If he has said that, is lying

[nocturne olive]

Add "good quality" and it's true

[young plover]

*an LLM as capable as neuro

[nocturne olive]

Not LLM 100%

[mighty thorn]

Bro Neuro is not a very capable LLM

[sage crag]

neuro is definitely using tools to control the model, but the question is about the intermediary interface

[mighty thorn]

Like not at all

[olive sable]

[mighty thorn]

Not in a bad way

It’s just fact

[sage crag]

to me its probable that there is a neural network that handles the movemenents since they dont seem hard-coded or procedural in nature

[woven quail]

ye, she's only there for entertaining purpose

[sage crag]

so i'd just assume that neuro is telling a neural network about what she wants the model to do

[woven quail]

don't try ask her for advice

[sage crag]

[crystal beacon]

i think only like a handfull are pre-coded

[sage crag]

https://cdn.discordapp.com/attachments/1187930764873056278/1438843835315654757/neuroModeMode.gif

[supple gulch]

When Vedal is fixing Neuro or Evil on Stream, do they get rebooted? Because it seems to me that they aren't rebooted/relaunched. But if a programm gets excuted and i change code, i need to relaunch it- right? Or does he work on the Memory (of the RAM) directly? if yes i'm more than impressed

[sage crag]

[olive sable]

obviously the claim from that one video about neuro being trained from twitch chat is false. the data is probably not fully ethical. but thats just how ai works.

[mighty thorn]

I doubt it’s just on chat either

Probably pretty similar to what I’m doing

[woven quail]

yes, they do, a several times happened

[olive sable]

i don't think you can really get the level of quality you would want without payimng thousands in author rights for data

[mighty thorn]

Just a different implementation I’d assume

[young plover]

Depends which part breaks. Some parts are probably separate microservices, and parts are unity which I think allows hot-reloading

[rough bloom]

nah, you can hot-reload stuff, there is a way to do that for Unity

[olive sable]

i doubt it was trained on twitch chat at all tbh

[mighty thorn]

Nah I believe it

[olive sable]

i don't

[woven quail]

surely he did not train her with /b/, right?

[olive sable]

i have no clue

it might have slipped in through the trainign data wherever he got it

[mighty thorn]

Def did

[woven quail]

well, that does explain why she's so unhinged

[supple gulch]

But the 3D today were like Poses that got connectect with each other (a bit like how its in a video game) and the stuff like the cart wheel were preanimated/fixed movemts, or could have Vedal train her that easy how to move in 3D and balance, I mean she fell down the stairs....

[old grail]

damn you guys got invaded by 3D neuro 💀

[olive sable]

?

[old grail]

this

[woven quail]

it's still about programming you know

[olive sable]

im pretty sure it was some type of model which outputs poses, and then he'd use interpolation between them

[old grail]

i just pointed that out

[old grail]

which*

[olive sable]

i already corrected the typo before you corrcted it lol

[old grail]

corrected*

not that time

[olive sable]

https://tenor.com/view/grammar-police-police-car-gif-16260565

bruh im esl

[mighty thorn]

Konii is a silly little specimen

[olive sable]

konii is a tv

[mighty thorn]

I honestly still don’t know what her pfp is

Besides just noise

[small sandal]

I felt that Vedal was surprised with the cartwheels

https://cdn.discordapp.com/attachments/1067638175478071307/1439340923388432696/msedge_SmidaR96fW.gif

[olive sable]

its a tv

[olive sable]

iirc picture of the subpixels with a filter over it

[mighty thorn]

Why have you progressed from 2.0.1 to e2025v3

[mighty thorn]

I think she left her ability to type

[olive sable]

oh god, konii spent the entire stream taking screenshots for reaction images

[small sandal]

https://cdn.discordapp.com/attachments/1067638175478071307/1439350940506394756/neuro_kungfu.gif

[mighty thorn]

She can only use mouse to send images

[supple gulch]

now my english isn't enough anymore XD

[olive sable]

?

you mean subpixels?

[supple gulch]

iirc picture

[olive sable]

its a photo

image

[amber fractal]

konii only want's to display 3d neuro instead of text

[olive sable]

understandable

[supple gulch]

a special type of photo/image?

[olive sable]

no

[sick owl]

Filians Minecraft practice is going well

[olive sable]

a picture is just a photo

synonym

[supple gulch]

"iirc" is than short for something?

[olive sable]

iirc its in the rules to speak english, but its a new chatter so ye

[olive sable]

if i remember correctly

iirc

[supple gulch]

thanks

[mighty thorn]

Which half of this image am I supposed to look at

[sick owl]

Yes

[olive sable]

left is fillians chat, right is an ai translation

[mighty thorn]

Ah I see

OOOH

That makes since

Given we have no way to translate between languages without a 24gb GPU, that makes perfect sense

[olive sable]

is that a chicken?

no

was that a chicken?

[quick quiver]

does Neuro use ElevenLabs to talk and if so how does she talk for so long?

[olive sable]

i think the answer is "we dont know"

to both

[sage crag]

[woven quail]

or, maybe it's bread

[olive sable]

that cant be a bread

bread served with lemon?

[woven quail]

yeah, i know, just refer to that one time

[olive sable]

based on the shape and lemon, that's a baked chicken

https://easyhealthllc.com/healthy-diet/diy-rotisserie-chicken/

its literally this

[sage crag]

🐸

[olive sable]

🐸

[sage crag]

at moderators

ban

frogsmog detected

[tender river]

shiro stellaris

[sage crag]

at shiro

[olive sable]

[sage crag]

bweee

[uneven frost]

there's this video that shows a model that can use text and song to make animations, but at the time of the vid it's said to be not real time...so unless vedal cooked or better techniques emerged in 6 months most likely: https://www.youtube.com/watch?v=7EA5JM1DI9Y&t=166s

[true hemlock]

portal 2 ran on NVS 310
a GPU about 29x slower than GTX 1650
on 2K resolution (2560*1600p)

game companies don't have excuses to make their game unoptimized if portal 2 with this graphics can run on such card fine.

CPU is i7 10700, with iGPU disabled

testing in i3 8100T gives same result

[hearty notch]

gm

[slate bear]

game companies don't have excuses to make their game unoptimized if portal 2 with this graphics can run on such card fine.
I'm agree

1. game company was Bound with contracts

2. game company wasn't happy if u play low end end spec

3. game devloper doesn't want optimize it like ark dev

4. unrealistic target

[olive sable]

well you see

they dont optemize it, because the time it takes to do so means the game will release later

[true hemlock]

game companies management sets unrealistic deadlines

so its not the devs fault

its the shitty management

[olive sable]

instead of releasing 1 game, they would much rather release 2 shit games and make twice as much profit

video game devs themselves, the coders, are very much so capable of optemizing it

[true hemlock]

also game companies want you to run their game at high specs so their game feels premium... as an excuse

[gleaming oxide]

Also, some people still buy/play games because they look good in 2025

Had a friend like that...

[slate bear]

now a days indie games devloper are achive more then big company ( i will sure that gta 6 will delayed any time soon )

[olive sable]

i dont remember what game it was. starfield or something. it re-rendered the depth stencil 4 different times each frame, because nobody ever bothered to change some of the engine settings to reuse the previous one.
its pure haste to make the game release in time due to shitty deadlines and management decisions

it sucks, but thats how the industry is rn

[gleaming oxide]

I kinda only play 2010's games and/or games that take inspiration on those early strong fundamentals

[slate bear]

MAYBE I WILL WAIT FOR PS6 OR BUY A NEW PC WITH USING EMULATOR

[gleaming oxide]

We have so much games nowadays that we have the choice to go back and play for example the first black ops instead of black ops 7

[olive sable]

eh

PS6 releasing wont make a difference

i doubt the ps6 will be anywhere near 4090 performance, and recent titles are already maxing out those

its a shitshow

[slate bear]

Gta 6 will fps locked on ps5

[olive sable]

well ye, and it will also be fps locked on ps6

i doubt it would do more than 60fps on the 6, nevermind the 5 trying to reach 30

[amber fractal]

insert gabe cube here

[slate bear]

Steam cube

[gleaming oxide]

PS6 ain't gonna have a whole lot of games

[olive sable]

the gabe cube is sadly only reaching ps5 levels of performance

[olive sable]

yep

[amber fractal]

I feel like a lot of performance is AI anyways

[olive sable]

or well, it will probably inheret ps4, and 5 titles, but in terms of new titles it will be even more bare than ps5

cuz they lost the exclusives

[amber fractal]

hardware isn't improving much besides being smarter for AI systems to run

[olive sable]

everything releases on pc now

[gleaming oxide]

we got the specs ? I didn't really dig deeper so Idk If it's gonna be worth

[olive sable]

we didnt

but

consoles are always behind in terms of tech

[amber fractal]

we didn't some people went data mining for similar

[true hemlock]

slightly less

[gleaming oxide]

Also, I'm wondering about modularity....

[olive sable]

especialy for a quarter the price of a gaming pc, you cant expect anywhere near 4090 perfomance

[true hemlock]

its an RDNA 3 with 28 CU

[olive sable]

we're talking ps6 now

[amber fractal]

hi Quack

[true hemlock]

17 TFLOPS but rdna 3 has shit schedulers so that pretty much is just comparable to an rx 6600

[amber fractal]

disregard this then, that is for the steam console

[olive sable]

unless they make the ps6 unreasonably expensive, its just gonna be like +30-40% of ps5 pro or something

[amber fractal]

I wouldn't put it past them due to ram prices alond

[olive sable]

yep

[true hemlock]

rdna 3 28 CU for gpu
zen 4 6 cores 12 threads
16gb ddr5

[olive sable]

we're talking ps6

ps6 specs arent out

[true hemlock]

bro was talking about steam machine

[olive sable]

i sont think at least

[amber fractal]

may as well answer the question

[gleaming oxide]

I was talking about gabe cube mb

[olive sable]

oh

[tender river]

i'm all for it make gamedevs target reasonable specs

[olive sable]

my bad

[true hemlock]

this is it!!

[amber fractal]

It was you who was off topic

[true hemlock]

since its practically comparable to a 7500f + rx 6600

[olive sable]

bwaa

[true hemlock]

(rx 6600 is considered budget level)

[olive sable]

i highly doubt the steam machine is gonna make triple A lower their performance targets, they will jsut say to use lower settings

[gleaming oxide]

Was my old setup

[true hemlock]

oh btw tested minecraft. 1.21.1, 12 chunk render, pure vanilla no optimization mods. 30 fps somehow? did mojang finally got up their asses and optimized the rendering engine?

on the nvs 310

which is in raw compute is 29x slower than a gtx 1650

[amber fractal]

1.20 had some changes to fix stuff up

[true hemlock]

ah

[olive sable]

i think they've been margninally improving it every now and then ye

[true hemlock]

nice

[amber fractal]

Still nothing compared to modded but it happens

[olive sable]

1.15 had a full back-end overhaul, and since then every couple updates had some stuff

[true hemlock]

i'll try sodium hold

[slate bear]

I care about bedrock player
The unplayable game rn

[true hemlock]

then i'll try shaders

[amber fractal]

Time to try nvidium on the nvidia card (real)

[olive sable]

bedrock int he latest update went form actually playable, back to how it was 5 years ago

[true hemlock]

oh yeah should try bedrock aswell since its C++

[amber fractal]

sam van typo

[gleaming oxide]

You reckon I can get Tekken 8 running smoothly on that ? Feel like it'll be good enough

[olive sable]

new bedrock update introduced numerous memory leaks, unreliable behaviour, glitches, lag, crashes

you name it

1.21 has been a major setback for bedrock

[amber fractal]

imagine basing on the mobile version

[true hemlock]

yeah sure

[gleaming oxide]

Had this old project of building an arcade cabinet from scratch but dropped it because It was kinda annoying to setup

Eh, maybe I'll get around doing it someday

[amber fractal]

Derailing a bit, but @gleaming oxide is a banger name

[gleaming oxide]

What entering NN without dampening his nape does to a mf

Appreciate it tho

[amber fractal]

Nah we derail all the time here

[true hemlock]

huh sodium doesn't give any performance improvement on the nvs 310

[amber fractal]

As I like to say, we are 50% on topic in #programming

[olive sable]

oh i didn't see that this message reacted to something

[olive sable]

not much better, if better at all

[amber fractal]

no optimizations for the nvs 310

[true hemlock]

nah its the exact same with or without sodium

[amber fractal]

It is time for shaders, what shaders tho?

[true hemlock]

tried photon

make a guess how many fps

[olive sable]

Sodium optimizes Minecraft by rewriting the rendering engine to use modern OpenGL features

well i see your issue, the nvs 310 is too old to not use old opengl

[amber fractal]

fps

[young plover]

pre-4.0 OpenGL is so ass

[gleaming oxide]

I have this very strange relationship with minecraft where, except for Hypixel Skyblock and the occasional mod I would try out back in the days. I never really modified my game. No mods, no shaders no texture packs...

[olive sable]

it hink this is OpenGL Version: 4.1 ngl

[gleaming oxide]

I kinda don't like the idea of modifying minecraft for some reason

[true hemlock]

2 fps, about the same as my pentium + hd2000 laptop running vanilla minecraft lmfao
also vram was starved so you have that black gap because it can't render more pixels

[amber fractal]

These days, it is better to give it a shot. But I respect the choice

[gleaming oxide]

Wasn't there earlier, is the goal killing your pc ?

[amber fractal]

darn, double my guess of 1

[true hemlock]

pushing nvs 310 to limit

(nvs 310 is about 29x slower than a gtx 1650)

48 cuda cores lmfao

[gleaming oxide]

PS3/XBOX360 minecraft got me in a chokehold

[slate bear]

It wasn't using 100 %
So we need modify windows driver side

[true hemlock]

oh sodium fully rendered. surprising

[amber fractal]

sodium is lighter generally

guess a few memory tricks still worked

[true hemlock]

i mean bsl

not sodium

oops

[olive sable]

im guessing sodium utilizes opengl 4.4, 4.5, sand 4.6 features to optimize it. you need opengl 4.3 for compute shaders for example

so that card isnt able to do compute shaders

[true hemlock]

uh no

nvs 310 has opengl 4.6

[olive sable]

are you sure about that?

i found other sources saying 4.1

actually

check what your pc says its using

[true hemlock]

https://www.nvidia.com/docs/IO/40049/NV-DS-NVS-310-May12-US-NV.pdf

hm

[olive sable]

that says 4.1

the thing is, they might have added 4.6 support in updates

[true hemlock]

mine support 4.5

[olive sable]

[true hemlock]

We only provide official support for graphics cards which have up-to-date drivers that are compatible with OpenGL 4.5 or newer. Most graphics cards released in the past 12 years will meet these requirements, including the following:

• sodium

[olive sable]

so its missing out on only the opengl 4.6 features

[amber fractal]

https://modrinth.com/mod/vulkanite-mod

[true hemlock]

i'll try bedrock

[amber fractal]

oh I recognize this mod creator, the nvidium dev

[olive sable]

afaik the only major difference then is mdi rendering cuz that got added in opengl 4.6

[amber fractal]

There is vulkan-mod but under my own testing it still sucks

[slate bear]

Watch this literally 2 core CPU with 5090 gpu
https://youtube.com/shorts/J6AnzGWUlBI?si=IJ2hD47DFXd0dXQ9

[amber fractal]

I remember posting on here with half the opengl fps

mainly because we miss a lot of optimizations that opengl did

[olive sable]

vulkan has less driver overhead, but is a lot more tedious to make actual optimization for

[slate bear]

Nivida is going good
Very weird combination with rtx
5090 with Intel dual core CPU

[olive sable]

eh

[amber fractal]

I can't find the vulkan screenshot so time to re run the tests

I'll start with the norm

I was about to crash out but I saw the fps cap

[true hemlock]

37 fps on bedrock

uh okay

[amber fractal]

opengl time, yell at me for 24gb ram all you want

[olive sable]

about 360 fps i guess

looking at the avg stat

[amber fractal]

The f3 screen just kills fps due to fetching all of the stats

[true hemlock]

didn't elvyn did some black magic witchcrafts on minecraft last time

[amber fractal]

we love mods forgetting to say sodium is a dep

[amber fractal]

I'm doing opengl / vulkan pipeline test

[true hemlock]

getting more fps with 248 chunk render vs my nvs 310 at 12 chunk render

[amber fractal]

which one of you is it

[olive sable]

1080p modded mc on a 3080, vs slightly less than 1080p on slightly less modded mc on an ancient gpu

[true hemlock]

i think she didn't even use distant horizons

just pure raw chunks

[olive sable]

ye

i dont like distant horizons anyways

bobby mod is where its at

[true hemlock]

last time she did distant horizons with shaders and got huge fps though hold on

[olive sable]

imo distant horizons looks bad, unless you set the real renderdistance high enough to not notice it

[true hemlock]

or use shaders

huh she could squeeze out 400-500 fps with shaders + distant horizons

[olive sable]

i can see the cutoff. idk.

[amber fractal]

distant horizons is fast when it works

[olive sable]

i woudl need to heavily tune the settings to a state where i wont notice it that much

otherwise it would bother me too much

[true hemlock]

@olive sable

[frosty locust]

https://www.youtube.com/watch?v=lUYQohnIQwE got neuro in vr chat lul

[olive sable]

thats getting there

i woudl still like the cutoff to be further, to a point where i wont notice the lack of textures

[true hemlock]

bro

i can't even see shit

[olive sable]

i can

[amber fractal]

I forgor to install vulkan

[true hemlock]

how are you even noticing the difference

holy shit

[olive sable]

its kinda my job

i do graphics

[slate bear]

I will set as pc wallpaper

[amber fractal]

Found it as well

[true hemlock]

@umbral thorn they stole it 🥀

[amber fractal]

I did take me a bit tbf

[olive sable]

for the far stuff, distant horizans looks good. for the closer stuff its okay-ish

[true hemlock]

should appreciate her more with squeezing out 38 fps at 248 raw chunk render

[slate bear]

MINECRAFT VIBRANT VISUAL has dealing with rendering system even change rendering system of java

[olive sable]

distant horizons is lod

not really a change in the render system

[true hemlock]

distant horizons didn't change shit

its just lod, just adds layers of lower quality but further render

[olive sable]

the rendering sytem is done by sodium

and nvidium or whatever else you pair with sodium

[olive sable]

ye that part is impressive yep

[slate bear]

Sorry it was vibrant visual

[olive sable]

thats not a mod

i myself would put normal render distance to 64 chunks, set bobby mod render distance to 128, ands then have distant horizons probably to like 512 chunks or something idk

i would have to paly around with it to actually test it out

[true hemlock]

bro even my device starts to struggles at 64 chunk

and that thing has a 13900

minecraft is mostly single threaded i don't fucking know how someone could pull off more fps than my 13900 with a fuckin 3000 series threadripper

i deadass was losing it when elvyn did it anyways

[olive sable]

bobby mod

[amber fractal]

man I love 30fps

[slate bear]

I suffer at 8 fps

I will try sodium

[true hemlock]

bobby mod didn't even optimize chunk rendering

[olive sable]

tis caching

[amber fractal]

I need to plug in the laptop holy

[amber fractal]

seconded

[true hemlock]

[olive sable]

i think you might be right about the 64 base chunks tho

it scales quadratically

[true hemlock]

yeah....

shit doesn't really make sense

248² is like

427x bigger

than 12²

but i dunno how

she's getting 38 fps

[olive sable]

probably 32 base, 128 bobby mod, and 512 distant horizons then

[true hemlock]

you sure?

the cached chunk by bobby is also rendered

[olive sable]

ye

[true hemlock]

so its going to act like raw chunks

[olive sable]

but there is less cpu side since you dont need to worry about chunk updates

[true hemlock]

fair

[olive sable]

so its going straight from vram to render

128 would probably fill most of the vram tho so id have to be carefull to not make it overflow into shared ram

[amber fractal]

I didn't want to update this pack for 1.21 for distant horizons so you get raw chunks

I still forgot to plug in the laptop

[true hemlock]

8 fps?

[amber fractal]

No shade on her, this is also insane for iGPU laptop at above 1080p

[amber fractal]

12 avg

[olive sable]

black ops 7 released. how???
it feels like a couple months ago that 6 was here

[amber fractal]

"Hello there, I like money"

[olive sable]

uses ai art apparently

70$ game

[amber fractal]

back down to 8 chunk for Quack, or 2

[slate bear]

READ AND WRITE IS STOPPED

Rip hhd

[olive sable]

[amber fractal]

30ish and 50ish in that order

this iGPU is suffering holy

but is also insane that 8->64 raw chunks only reduced by 1/3

[rigid snow]

minicar ft game ing

[amber fractal]

[rigid snow]

have y’all seen the opencl worldgen thing

it’s actually crazy

it generates chunks faster than they load from disk usually

[amber fractal]

no I haven't actually

[olive sable]

[rigid snow]

https://youtu.be/C1fpUZ4o2Jc

found it

[amber fractal]

It seems like it is worth updating my client to 1.21 after all

I've been on 1.20.4 forever now

[rigid snow]

C2ME test build with gpu accelerated worldgen using openCL

[amber fractal]

Not due to this video, but the surrounding mods

like shaders functioning in voxy

which I still prefer over distant horizons

[rigid snow]

you can use c2me still it’s pretty good

[amber fractal]

I already do

I also use noisium on top of it

[patent shard]

what the fuck, I think the battery is toast
what the fuck kind of scam is this.. you pay more to get your battery bricked?
gets a little too low charge and the battery is just trashed? wat..

why would you get this horrible battery over a lead acid one.. at least that won't completely brick itself if the battery gets low charge

[amber fractal]

test world seems fricked

[olive sable]

gel batteries are lead acid

but they're fully enclosed

and the elctrolyte is in gel

hence the name

[slate bear]

I will back after play minecraft

[amber fractal]

I also need to update java as I do not have java 22 for c2me in this version

[olive sable]

arent gel batteries chepaer btw?

[patent shard]

I would hope so, I thought the opposite

all I know is they brick themselves if the charge gets even a little low

[olive sable]

Gel batteries should ideally not be discharged below 50% of their capacity to maximize their lifespan.
i guess you completely discharged it?

Discharging below this threshold accelerates sulfation (hardening of lead sulfate crystals on the plates), which is often irreversible and causes a permanent loss of capacity.

i see

[patent shard]

it was def not completely
yeah, it's apparently very bad

[amber fractal]

ohhhh that is a 15gb update

I guess it has been a bit

@olive sable surely no errors here

[olive sable]

fog

is this a water or chunk glitch?

or both?

[amber fractal]

Before I say not distant horizons but same class of mod

[olive sable]

[amber fractal]

This is just out of the normal render distance and enters voxys

[olive sable]

[rigid snow]

why it look like that tho

[amber fractal]

for 8 render distance it is not bad

[amber fractal]

classic voxy bug

motion blur :RAGEY:

I need a better screenshot tool for games

there we go, no blur

[rigid snow]

i go out of my way to add a motion blur shader in minecraft it makes it feel very nice

idk motion blur is cool

[olive sable]

blasphemy

motion blur has its place ngl

[amber fractal]

the blur was added by the screenshot program

[olive sable]

as long as its done tastefully

[rigid snow]

especially if i’m playing pvp or something

[amber fractal]

I can see the vision

[rigid snow]

makes me less lost

[amber fractal]

barely (lol)

[rigid snow]

it’s grounding

idfk

just something i do

feels nice

[olive sable]

vignette motion blur, to make sideview look faster is good. gives a feelign of speed

[rigid snow]

especially on like 400 fps it’s not bad

[olive sable]

[amber fractal]

If you want to spin up a client @olive sable besides water being white for some reason, I think you'll like the rest of it compared to distant horizons. Trees have textures

[olive sable]

honestly, not rn

[rigid snow]

yeah voxy is better

[amber fractal]

The apporach is nicer

[olive sable]

normal mc already takes up 10 min to launch due to fuckign microsoft account shit

[rigid snow]

i don’t like it in theory but ends up working

[olive sable]

nevermind having to do the curseforge stuff

i cant be bothered

[rigid snow]

GET A THIRD PARTY LAUNCHER SAM

do not use m$ or curseforge shit

[amber fractal]

I'll hand this to mlnt

[olive sable]

you still need to log in to microsoft account via curseforge launcher

[rigid snow]

once though

[olive sable]

nop

every time

[rigid snow]

weird

i beg you just use prism launcher

you’ll never look back

[olive sable]

how am i suppsoed to use prism when i use curseforge?

[tender river]

i use nix as my minecraft launcher

https://gist.github.com/eyJhb/623c723ddf068a1c8de26e1ca467f002/

[olive sable]

thanks for your input leaf

[tender river]

ye ye important

anyway prism is flexible probably supports curseforge

yee it does

also modrinth

[rigid snow]

what does this mean
prism has curseforge modpack support built in, like a browser and all

[olive sable]

cool i guess

still cant be bothered to set it up rn

[slate bear]

My pc running minecraft on 8 to 14 fps

[amber fractal]

[rigid snow]

literally just install it from aur

or maybe it’s in the main repo

i have 0 idea

[olive sable]

i would still need to port everythign from curseforge to prism

too much work

[tender river]

ye paru -S prismlauncher

in main repo

[umbral thorn]

https://cdn.discordapp.com/attachments/1324872678791975004/1439388620610867300/GIF_20251116_064421_186.gif

[tender river]

true true

[olive sable]

if i cant be bothered to spend 10 min launching curseforge, what makes you think i can bother spending 15 min setting up a different launcher?

[umbral thorn]

neuro was so cutee

[amber fractal]

Why did I even bother including distant horizions

[rigid snow]

it does not take 15 minutes

[olive sable]

sure

[tender river]

https://cdn.discordapp.com/attachments/1067638175478071307/1439352964455075880/neurospin.gif

[olive sable]

🐸

[umbral thorn]

[tender river]

model so good shading so good

[olive sable]

yep

[tender river]

and neuro has so much more she can do now

[olive sable]

i was impresses

[tender river]

great stream

[olive sable]

[tender river]

like if you compare it to her live2d its very similar despite being 3d

[olive sable]

cell shading

[tender river]

https://imgur.com/a/NJHvyBH

[olive sable]

i think at least

or just good texturing

[umbral thorn]

ooooh

[rigid snow]

i don't think it's even cel shading i didn't see any shadows i think

nvm

cel shading

[olive sable]

the neck

but im not sure if its baked yet

i think its baked into the texture???

the line-art is done as a shader, but i think the shadows are texture based. don't quote me on that tho

[amber fractal]

For the chat, just to finish off this autism moment, sunset

[mighty thorn]

Hey I know you

[umbral thorn]

https://cdn.discordapp.com/attachments/1324872678791975004/1439388620610867300/GIF_20251116_064421_186.gif

[molten ore]

The pose to me looks like someone who is sitting on a chair backwards

[patent shard]

wat

[neat cosmos]

ofc-

[rigid snow]

classic ue slop moment

[short ingot]

chat any ideas why this did not work?

# static/.parsie/modules/autoload.py

from src.core.types import ASTNode

class AutoloadSymbolProvider:
    def __init__(self, pm):
        self.pm = pm  # This is ParsiePluginManager

    def tokens_to_ast(self, tokens):
        """
        Syntax:
            autoload [PluginName]
        """
        if len(tokens) < 2:
            return ASTNode(kind="autoload", data={
                "error": "missing plugin name",
                "raw": " ".join(tokens)
            })

        raw = tokens[1].strip()

        # Expect format: [Outerlands]
        if raw.startswith("[") and raw.endswith("]"):
            plugin_name = raw[1:-1]
        else:
            plugin_name = raw

        return ASTNode(kind="autoload", data={"plugin": plugin_name})


class AutoloadExecutor:
    def __init__(self, pm):
        self.pm = pm  # This is ParsiePluginManager

    def can_handle(self, node):
        return node.kind == "autoload"

    def execute(self, node):
        plugin = node.data.get("plugin")

        # The plugin file path for Parsie
        file_path = f"static/.parsie/modules/{plugin.lower()}.py"

        try:
            # Force import early
            self.pm._import_file(self.pm.pm._abs(file_path))
        except Exception as e:
            print(f"[autoload] failed to load '{plugin}': {e}")


def create_symbol_provider(pm):
    return AutoloadSymbolProvider(pm)

def create_executor(pm):
    return AutoloadExecutor(pm)

def register(pm, module_key):
    pm.registry.register_symbol(
        symbol_name="autoload",
        module_path=module_key,
        factory_name="create_symbol_provider"
    )

    pm.registry.register_executor(
        kind="autoload",
        module_path=module_key,
        factory_name="create_executor"
    )

[short ingot]

what happened to your fps counter???

was the fps changing mid photograph?

I just realized-

[amber fractal]

the screenshot tool doesn't actually take the screenshot while it is frozen in place, so it both fades back to normal and captures the screen

[short ingot]

oooh

why'd that take so long to type

[amber fractal]

Same to you, why are you instant

[short ingot]

uhmmm idk

[short ingot]

I just type fast, ig

is my code okay actually

[amber fractal]

Nope, I'm going back to watching anime for the night. I just came to answer that question

[short ingot]

kay!

[slate bear]

I have questions in my mind why it has error checking?

[short ingot]

wdym?

[slate bear]

Just tell me

[short ingot]

I dont understand what you mean

what error checking?

[slate bear]

Don't be a stupid

print(f"[autoload] failed to load '{plugin}': {e}")

def create_symbol_provider(pm):
return AutoloadSymbolProvider(pm)

def create_executor(pm):
return AutoloadExecutor(pm)

def register(pm, module_key):
pm.registry.register_symbol(
symbol_name="autoload",
module_path=module_key,
factory_name="create_symbol_provider"
)

[short ingot]

...

what?

what's wrong with that?

[jagged turtle]

is there something wrong with error checking?

[slate bear]

NO

[obsidian mantle]

I smell i

[slate bear]

I'm just asking why do u need error checking

[obsidian mantle]

[slate bear]

Do u cheak the code

[obsidian mantle]

Not yet its too long

Have to learn some kanji first today

[slate bear]

Can u also teach me bro please

[obsidian mantle]

Anki+kanjidamage deck

[slate bear]

I REALLY WANT LEARN JAPANESE

[obsidian mantle]

Learn hiragana on tofugu first

[short ingot]

from my experience, nothing had gone wrong with print(f"[autoload] failed to load '{plugin}': {e}")

yet

[jagged turtle]

"dev" btw

[jagged turtle]

"NOOOOO BUT YOU CANT DO THAT YOU ARENT SUPPOSED TO EXPECT ERRORS YOU'RE SUPPOSED TO HOPE AND PRAY" :Smadge:

/s

[short ingot]

are you saying you can debug this dump?

its very detailed

but a core dump would be better

at most

[hard raptor]

@rigid snow @fast pagoda i would like both of your opinions on this

[short ingot]

idk how to get codespaces to dump python's memory yet

[slate bear]

Dam it

[hard raptor]

Maybe slight headphone warning if your headphones or earbuds or whatever have like boosted bass

[slate bear]

I thought that it was a sample

[hard raptor]

it def is not

[slate bear]

Where did I get full file bro

[short ingot]

waht

I dont understand what you mean

I'm not in your head

I cant decode your texts

[slate bear]

Give ure files

[short ingot]

are you gonna steal it or smth?

[short ingot]

already sent it here

[slate bear]

No it just for cheak

[short ingot]

cheating?

checking?

i, wdym?

[slate bear]

I don't have modules

[short ingot]

ill send u the repository gimme a moment

https://github.com/nakurity/nakurity-language/tree/prototype/dev-1

[jagged turtle]

?

[short ingot]

you know like, a core dump?

[slate bear]

I hate to read licence anyways

[jagged turtle]

I'm pretty sure there's a verbose mode you can run python in but idr if that's true

[short ingot]

oooh

[slate bear]

A minor error

[nocturne olive]

What is this wizardry?

[rigid snow]

i do not know what to say

i'm not speechless

i just don't know

[hard raptor]

just a lot of knobs being turned automatically

[nocturne olive]

Complete wizardry
Instrumental stuff is pure black magic

[hard raptor]

the Chorus is doing most of the heavy lifting to create the wierd harmonics

i hope this give a bit of an idea of that

[hard raptor]

What do we think about this?

[slate bear]

I will cheak after I will home

[slate bear]

Nice project

[obsidian mantle]

windows doesnt let me to even make file with " in its name

[nocturne olive]

Windows bad

[obsidian mantle]

whats a weird name .. print ("aaa").py

i guess it should have been aaa.py?

cand find "aaa" being anywhere in the source

[short ingot]

that code is supposed to be ran as python code

its an outerlands module

its alien code

[obsidian mantle]

it tries to open file that cannot exist on windows

[short ingot]

outerlands module is supposed to resolve and interrupt that from being imported as a file

https://github.com/nakurity/nakurity-language/blob/prototype/dev-1/workspace/static/.parsie/modules/outerlands.py

this line fired

but I assume something is wrong with the interrupt system in parser

wait, in vscode can you connect to a codespace with others?

can I share my codespace?

@jagged turtle do you think its possible?

[obsidian mantle]

this is where it went off rails i think

[short ingot]

yea

[obsidian mantle]

i guess it was supposed to execute "print" but instead it sent it as a filename to someone

[short ingot]

something is either wrong with check_and_capture()

or the interrupt system in parser

[short ingot]

its supposed to cache it to an variable, until it finds the finished definition line

because the interrupt system is an unimplemented multi-line definition support

[obsidian mantle]

is github search just cosmetic

cant find things which i stare at

[jagged turtle]

https://marketplace.visualstudio.com/items?itemName=MS-vsliveshare.vsliveshare almost got it off by memory

[short ingot]

tysm!

sorry for the ping

btw

@obsidian mantle are you able to connect to this? https://prod.liveshare.vsengsaas.visualstudio.com/join?F3BE75363340A0ACBC92CE3A8CEC978C6BE0

[jagged turtle]

thx 4 the link xD

[short ingot]

Np!

I'm happy to share all sensitive information whenever xD

[obsidian mantle]

hmm

oh it worked through browser

[short ingot]

I like the anom who joined and I gave him read-write thinking it was u xD

its ok! they will be nice, right...?

[obsidian mantle]

i typed "vituha" when they asked for my name

[jagged turtle]

[obsidian mantle]

oh wait it says im anonymous and readonly

this is cool wtf

[jagged turtle]

man

can't wait to get neuropilot working with it

[short ingot]

yea!

[short ingot]

lol was that anom, you?!

[jagged turtle]

[short ingot]

uuuhh, anom 1 has been ejected...
he was not the imposter

[jagged turtle]

I feel like rpc coordination is gonna be very annoying to deal with for this

[short ingot]

did you like how in my code. we had to import commenting?

[jagged turtle]

import w h a t

[obsidian mantle]

i think we want to know where "check_and_capture" exits

it just tells it returned nothing

but we dont know what it did

[sage crag]

[short ingot]

its in outerlands.py

at static/.parsie/modules/outerlands.py

[short ingot]

Commenting on the code.

[short ingot]

without this, the program crashes

[obsidian mantle]

yeye i see the function

i dont understand what it should do

and what it actually does in our example

it seems like it doesnt do what its supposed to do

add prints before every return in it

[short ingot]

its supposed to append lines after outerlands boarding [below]
to a variable, a string.
and to append lines until it reaches: [finished-definition]

this is the test file for it

very basic

[obsidian mantle]

but after this function, plugin manager gets some weird filepath

[short ingot]

I'ma take it a step further. requiring you to download a package to recongize comments

[obsidian mantle]

its either the function itself or somewhere who is supposed to process the result

[short ingot]

its not supposed to get that

outerlands is supposed to process the result

it should've interrupted the parser from resolving a file path for it

https://github.com/nakurity/nakurity-language/blob/prototype/dev-1/workspace/static/.masha/plugins/parser.py#L197

the interrupt system is either wrong