#Neuro ARG - Cracking the "Numbers" Video
1 messages Β· Page 4 of 1
It might be SSTV, just as well
fug
didnt I post a spectrograph of numbers already
i don't think anythings in the spectrogram but if it helps at all the chord progression is a VI-i-vii-VI-i-v (6-1-5-6-1-7)
what about study?
I did that too iirc
theres technically two bars of the 6 in both instances so (6-6-1-5-6-6-1-7)
honestly i highly doubt this means anything since its one of the most common chord progressions
Seems like nothing
this is with melodic range spectrogram
you got a 4k monitor?
the drum patterns and instrumentation likely also mean nothing bc it legit sounds like the most generic trap beat you get when you search "royalty free trap beat"
Yeah
i feel like if there was something itd sound weird
but it just sounds normal
idk thats my 3 brain cells speaking
holy
numbers with more horizontal resolution
i love this
study with more horizontal resolution
those large dots are just the bassline(chord progression) i just mentioned, so its not like morse or something
right but it can be a code
.
have we tried reversing the audio on both tracks?
idk
i beleive so
Anyway i am off to work, good luck with arg
π
gl guys ill go mimir
@pale fossil has leveled up! (3 β 4)
how does one get embed perms?
check FAQ
what has happened since yesterdya
someone pointed out something between 31:10 and 31:11 in the study video
its another hash
its been solved
damn
did that program get made? if so i will run
can you be more specific
the one to try to brute force an AES that is 17xxxxxxxxxxxx24
new pfp 
What if it's like the fnaf books where each age range has its own "body"
oh I posted this in the wrong thread lol
meant to post it in #1112082183235047506 because they were talking about it
lmao
It is possible we got a hit but just don't know it since it only checked for UTF-8
true
this got decoded i think
this thread is now the place where we talk about the ARG without vedal seeing us.
admins can tell vedl
Would it be any useful to note that the animation used in the Numbers vid is hiyori_m10.motion3.json?

jajajajaja
I need to figure out why my code is still allocating in the heap π
uh, one minute
its not
i have my theory
the hash is base64, but we're not meant to decode it entirely
theres something about "17" is first
which could mean we have to cut it 17 times
and if you notice almost every number from 1-10 is mentioned in the song
so if we follow the song and order them somehow
or somehow substitute something in them
we can decode it
1388 is not divisible by 17
meaning theres some things we have to do to the hash if this theory is true
i've already spotted one issue with this but that's easy to fix
The ARG is designed for us to find the key under the mat, but we try to poke the keyhole with a fork until the door opens
because theres 236 different mats depending on how you read the instructions
The ARG is designed for us to find the key under the mat, but we try to poke the keyhole with a fork until the door opens
alright time to change crypto library in hopes of speed
Agreed
and theres also 19 different keyholes
if we found the mat that we know is the right one, then its 1/19 chances
instead of 1/4,484 chances
except we're impatient/lazy so we doing 1/(19*1,000,000,000,000)
that guy...
it's the irrationally that comes with being stuck.
We have already figured out that the 2 to start with is on the 129 character
is the 129 symbol*
looking at the "numbers" video, i have no clue how one would flip the numbers
but maybe im looking at the wrong place
2 is the 130th, after 129 other characters
Yes, I was wrong. 130th
So what else do we have analyzed
I'm checking the relationship between the first 2 and other numbers
try to get all of the numbers
I don't even see what the reason is to go to point X then point Y and point Z
I think its dumb
why does it matter where the 2 is?
Because the hash is not correct and if we edit it according to the hints in song, it can be decrypted
Just a little hint, base64 should be divisible by 4, maybe we should keep this in mind when editing.
hopes of speed have made it 2.5x slower
rip
Is this code even theoretically capable of doing anything?
i don't even know what i'm looking for
people are lost so i thought brute forcing might be a mildly viable option
suprise suprise the silly turtle person made this difficult
You're right there
not bad
damn, capped at 350m / sec for some reason
2m/1 billion
I can help you
Give me the script and tell me how to run it
yes
Just for fun
install the rust toolchain, git clone https://github.com/owobred/neuro-brute/tree/png (make sure you're on the png branch), set RUSTFLAGS = "-C target-cpu=native" as a env var, then cargo run --release in the root of the folder
rust toolchain is from https://www.rust-lang.org/learn/get-started
Already installed
oh you will want to edit the number of threads it uses because I'm still lazy and don't work it out dynamically
I see
line 20 of src/main.rs, change the number to the amount of threads you can use effectively
im trying to learn how to use rust it looks like a nice language
It isnt
it is
DM
they might be closed here
i want to see the reasoning as well
@tiny garden can you cut the hash in 3 parts
and run it thru base 64 decoder separately
like
its base64
the hash is just changed and the steps to make it normal are in the video
it kinda makes sense
which hash
are you serious
im not even gonna tell you
you should know
the one you've been bruteforcing this entire time
thats the first one discovered
its base64
base64 isn't a hash lol
an encoding is the usual name
that then
I guess that could be something
idk I won't really be able to do much for a while, gonna be up on the hills in like half an hour
alr
Hm
Hm
I made something that can generate numbers from the lyrics but I can't figure out how to use this crypto library
could someone that has a working decryption thing try these numbers as the key please?
[176, 17624, 17600675, 17606750, 17933675, 171237182, 175460675, 1724600669, 1724606744, 1724933669, 1746571820, 1753256750, 1758657182, 1759325675, 1760066924, 1760674424, 1760972153, 1769721530, 1793366924, 1793372153, 17102372153, 17123717624, 17123721872, 17241237176, 17245460669, 17546066924, 17546972153, 17600721485, 17607214850, 17933721485, 171237215282, 171405572153, 172446571814, 172453256744, 172458657176, 172459325669, 172460972147, 172469721524, 172493372147, 174533721530, 174657181424, 174837218720, 175325674424, 175343721530, 175460721485, 175557218720, 175853372153, 175865717624, 175883721872, 175932566924, 175934372153, 175955721872, 176097214724, 176972152424, 179337214724, 1710237214724, 1712372186624, 1724102372147, 1724123721866, 1724546972147, 1724600721479, 1724607214844, 1724933721479, 1746572152820, 1749155721530, 1753257214850, 1754555721872, 1754697214724, 1758657215282, 1758915572153, 1759325721485, 1760072147924, 1760721484424, 1793372147924, 17123721527624, 17140557214724, 17241237215276, 17241405572147, 17244533721524, 17244837218714, 17245343721524, 17245460721479, 17245557218714, 17245853372147, 17245883721866, 17245934372147, 17245955721866, 17453372152424, 17483721871424, 17534372152424, 17546072147924, 17555721871424, 17585337214724, 17588372186624, 17593437214724, 17595572186624, 172446572152814, 172449155721524, 172453257214844, 172454555721866, 172458657215276, 172458915572147, 172459325721479, 174657215281424, 174915572152424, 175325721484424, 175455572186624, 175865721527624, 175891557214724, 175932572147924]
ranges: [0..62500000000, 62500000000..125000000000, 125000000000..187500000000, 187500000000..250000000000, 250000000000..312500000000, 312500000000..375000000000, 375000000000..437500000000, 437500000000..500000000000, 500000000000..562500000000, 562500000000..625000000000, 625000000000..687500000000, 687500000000..750000000000, 750000000000..812500000000, 812500000000..875000000000, 875000000000..937500000000, 937500000000..1000000000000]
We took the easy way\
not sure what that is but I'm gonna guess it didn't succeed
oh
you're brute forcing the ciphertext and assigning those ranges to different ppl?
We're checking if result are UTF-8, if text UTF-8 then saving result
yo we brute forcing again ? nice
For files we are checking first few bytes if its match example
theory is we might get an url to something like a new clue video
On more powerful CPUs
I think so, too
or we might get a phone number
I live in mother Russia, the call will be expensive
it would be for this
so if the numbers video give use a number we can try sending a messageto a phone and recieve another clue maybe ?
OK WHAT THE ACTUAL FUCK IS HAPPENING HERE
I'm not very good at rust. And guy who wrote the code ran out of mobile internet
So we're waiting for him to do it
hi quax people are doing silly stuff here
What the actual fuck is this
people got to far on this
clue from the other video, it's code to send message to a phone
0 progress
ah ok
here is what everyone is doing
https://drive.google.com/drive/folders/1qm6GSRlstTmjbyOFCuH1OhSmPK4pi4Aa?usp=sharing
https://docs.google.com/document/d/1n3nEfer7vQcE4yIQAQyoBiWsi3r7f5mfgwmEUNXZ3AI/edit seems like a dead-end for now
2023-05-30T08:49:55.441753Z INFO neuro_brute: 169,990,000,000/1,000,000,000,000 (16.999%) current rate 45,300,000 keys/second
Code isnt optimized, speed is too low
NAH BRO HOWD YALL KNOW ALL OF THIS
ICANT
x)
Static-sama is trolling and I'm tired of telling people that
every clue yes but from official sources
@tiny garden tried your PNG branch, I get 450m/sec and then it crashes due to stack overflow
I'm thinking that once we can decrypt the Numbers description we can use the same key to decrypt the phone number there as well
Delete this if
BTW, we have already processed the png
no hits?
where tokio
I ran that one without any hits
smh
Poor optimization makes it slow
I provided the server for the bruteforce
@heavy tangle has leveled up! (3 β 4)
I'm also doing the same, since my server is sitting idle rn, we should probably run different attempts though π
Agreed
WHAT
I can give it more, but is it supposed to consume that much?
I ran directly in PVE, so as not to virtualize. Didn't see how much it consumes, so I don't know
might be because I'm using cargo +nightly, doesn't seem to build on stable
#![feature]may not be used on the stable release channel
It should be nightly
2023-05-30T09:48:35.484114Z INFO neuro_brute: 329,112,250,000/1,000,000,000,000 (32.911%) current rate 45,375,000 keys/second
It sooooo boring
which commit are you on?
Looked at the consumption graph
It doesn't have to be that way. It consumed very little for me. I can't be exact, but in general 0.5-4gb
The error is large because of the VM
what consumption graph?
On PVE manager
I'm just using Hyper-V, this is a windows machine
I don't think it's something too complicated. The clue probably something on the surface, but we can't see it because focusing on something else.
The second song related to the first song, and it seems like the second song is a hint for the first song. People managed to solve quite a lot on the second song compared to the first one.
@heavy tangle Fixed it by reducing the number of threads from 64 to 32, getting a rate of 310m keys/s now
Could be more if I can use more threads
actually it wasn't the number of threads, it was the target-cpu=native that was causing it
getting 480m keys/s now
480m/s should be like 30 min ?
yeah, but this one was already run without any hits
I'm just running it for fun, I don't expect results
whatever ill find out the result the intended way
was hoping someome was smart enough
its an unlisted video
brute force unlisted videos ?
@minor fiber has leveled up! (6 β 7)
Stop
You need to change something in code
Since it's not a png, you need to look for the presence of a jpg
that's going to be somewhat difficult, a jpg doesn't have a well-defined magic header
change the header check to match the first 3 bytes of the jpeg header https://en.m.wikipedia.org/wiki/JPEG_File_Interchange_Format
not all jpegs have a JFIF header
*hex
random info dump
16, '1', 26, 'Q', 17, 'W', 11, 'o', 19, 'q', 21, 'Z', 21, 't', 18, 'C', 27, 'd', 21, 'r', 25, 'w', 18, 'b', 19, 'i', 21, '/', 21, 'D', 22, 'l', 16, 'k', 11, 'K', 23, '4', 20, 'I', 19, 'f', 16, '=', 1]```
I'll use just the SOS, I'll prob get false positives, but w/e
how many times the letters appear
btw, something else that's worth a shot is converting the numbers to letters before using it as a key given the "numbers matching the letters" lyric
i got video about schizofrenia as recommended on the numbers vireo
True
that means we are getting closer
i think

string = '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'
startingPoint = 0
while startingPoint != len(string):
b64string = ''
for i in range(startingPoint, len(string)):
b64string = b64string + string[i]
encoded = b64string.encode()
try:
decoded = base64.b64decode(encoded)
if '\\' not in str(decoded):
print(decoded)
except: pass
startingPoint += 1
not cut in any way
How are my favorite schizo people doing
agony
I'll try and rewrite that section so it doesn't try and use some stupid amount of ram when I get back
fixed it, target-cpu=native caused it
how do you have interwebs?!
anyway, no hits searching for jpegs either
probably not worth throwing any more compute at it for now
anyways imma get walking again
Do you guys think this whole arg is just for people to get used to the new Neuro voice
Undetected, unexpected
Wings of glory, tell their story
i dont think the voice in the video is neuro
thats sabaton not subathon
It's Neuro, too similar to be not neuro
x)
To natural, to be neuro TTS
Well her tts need an upgrade tbh, it's too robotic compared to her vtuber model
We don't know yet and the voice can be easily edited, so it's too early to say for sure that it's Neoru. Although I admit I can hear her somewhere in the distance in that voice too
he's working on it
V2 voice coming soonβ’οΈ
I imagine if new voice is the first thing to do, perform one of these songs, but that doesn't fit in the video a bit then no?
/ cannot be generated in base64, so we need to process each string separately.
You have unlocked new role
You have unlocked new role
Use this as a clue and process the hash
CLUE 
OMG, ITS STATIC SAMA?????
(also you can post images now)
Finally
?
Nevermind
I said the wrong thing
Anyway
Regardless of the resolution, each line begins with /
I think it's something
it is not. youtube automatically makes breaks on symbols like that
Shit
you can put the thing in notepad and stretch it around to get different line breaks, woo, truly a clue
when i paste in notepad it format like that
That's too bad. I didn't know about this feature
its pretty funny tbh, i had abit of fun following along
Sure it's fun, but there's no progress
must be a clue
well i am personally not very good with numbers and stuff in here
What is the clue?
nah nothing
we discover a good lead last night which is the 130th character in the hash is a 2, matching the two lyrics about counting days and start with number 2
anybody followed down that lead?
it's possible that we have to alter something in the hash itself before decoding it
this is also a possibility that we ignored because we rushed to decode it first
can someone send me a link of a site that can be used to decode base64? I need to test some keys
oh wait me counting the days was actually useful lol what about the 1432 days ?
The Cyber Swiss Army Knife - a web app for encryption, encoding, compression and data analysis
unfortunately there are only 1388 characters in the hash
thanks
Did somebody follow my trail that 17 and 24 are maybe 1597 and 13 because of fibonacci ?
unlikely
Is the Line "Add the numbers 2 4" mean 6 or 24
no idea
idk
Finding all these numbers
Start with number 2 yeah
Matching all the letters
572943
Match all the letters with 2572943:
a=2
b=5
c=7
d=2
e=9
f=4
g=3
Add another 9 yeah (watch as the same numbers between 2-9 is retained which can be matched at the end of the last line of the lyrics):
2572943+9=2572952
maybe the final key will be straight up only letters which will be matched based on the final manipulation of letters
So the fibonacci is a false trail in that specific time frame from "Add another 6 yeah" to the last "ABCDEFG" ?
I thought because it was shown there that we needed to run the numbers through fibonacci
in any case we need to think how can the lyrics help us switch stuff in the encoded string
cuz i think thats what it is
the base64 string is switched
and the lyrics helps us fix it
might also be start with the 2 so 572 +9 = 581 -> 581943
give me ideas on how to do this, any
basically any ideas
@stoic yacht @spring widget wanna base yourselves around that idea and help?
The Base64 string is modified, and the lyrics tells us what to substitute in order to decode it.
Darkeew I am absolutely clueless on anything that has been happening in this thread
yeah but that wouldnt make any sense, since if the purpose of:
Finding all these numbers
Start with number 2 yeah
Matching all the letters
572943
is to establish a value for the letters of ABCDEFG, and since 1 is not included, then that means any solution with 1 or any number that doesnt exist on the established value of the letters is wrong
@sacred ruin this is one of them good lead to start with
I have no idea what "base64" means, I am actually useless here
we're beginning a new investigation
the description is a clue
but that clue was modified
could also be 5729439 - > abcdefg
and the lyrics tells us how to fix it

Don't forget that this could be the key
basically think of it as a normal phrase
no it can't be
and use the lyrics to fix the phrase
nah
anyone has other ideas?
okay darkeww I'm not that clueless, but like as soon as people start talking about bytes you lose me
i'm not talking about your idea
forget the bytes
Why not
ill treat you like a child dw
ah- ok
im saying that this is not possible
smartest one here
According to ASCII (American Standard Code for Information Interchange), the numerical value of a new line is 10
help us solve
have we attachted values to abcdefg or are the letters mea nothing
yes
im gonna have lunch
Well I can't help right now because We're making Neuro play with the demon core on the event canvas
yummy radiation
This is beautiful
don't get sidetrack now
start with the 130th (the 2) character in the hash
it matches with the lyrics
start with uhhhh
start
just start
do anything
too many people are getting trolled with static
so we have basically no one solving the actual arg

25+ people on the static sama doc 
and making no actual progress
Bruteforcehttps://cdn.discordapp.com/emojis/1072901352902103100.webp?size=48&quality=lossless&size=40
the guy behing that discord alt really feeling good with all that attention
If you encrypt text in this site, it gives you an output that looks mostly like base64 but has slashes in it https://www.devglan.com/online-tools/text-encryption-decryption
So that's a likely candidate imo
at least it made us think to us something else than AES
Wow, SAL caution
SAL?
Security bot is checking every link
oshit a couple of those numbers from before reported a success let me try it
Any bruteforce attempt should score the putative plaintexts by entropy instead of magic bytes to avoid false negatives+positives and be payload agnostic
nah neither worked I guess my script has a bug
holy shit I do not understand half the words in this sentence, i kneel
Script is checking for valid UTF text
basically instead of checking if the text don't have corrupted character (which is what we were doing) they are saying to check how random the text is
Basically instead of trying to see if it decodes into text or an image we should instead look at how "random" the bits are
Is there an ELI5? :p
Aye but that'll be wasted computation if the payload turns out to be something other than printable plaintext
Anyway before a brute force attempt is actually going to be successful we need better clues about what the key should look like
Because the current code makes a lot of assumptions about the key
In fact it doesn't matter if there is a troll or not, we have no idea about it
do we know what cipher is used

at some point I should make it keep track of its best finds
might not even be any keys if it's just Base64 but scrambled
Any AES decrypts with an incorrect key should give a "plaintext" indistinguishable from random noise meaning it'll have a high entropy. If you get a valid key during the brutefroce it'll give the actual payload. Unless the actual payload is encrypted further (With modern crypto) it'll have a lower entropy. It's a more robust method of checking if your key gave you the proper plaintext or not instead of checking for the prescence of magic bytes e.g. ff d8 for JPGs, or percentage of printable characters
^ but if you're being a nice arg creator it shouldn't be encrypted 2 times because then you're just being weird
I could probably use something like https://docs.rs/entropy/latest/entropy/
API documentation for the Rust entropy crate.
but if you check only a small part of the decrypted wouldn't it change the entropy ?
it is entirely possible that the base64 snippet in the SMS-program is a phonenumber, given it's placement in the code. But what format of phonenumber i do not know
we do not, the only reason we're assuming 128-bit AES is because the base64 decodes into a multiple of 16-bytes
the data seems too long to be just a phone number
i think he is speaking of DQ5zl4ighWwiag7y+cWFQg==
ye, i just realized it wasn't the numbers video that appears. I'm a confused person :p
yeah I was talking about that one too
you think it's too long ? i would have guessed it's too short
well i tried giving numbers some thought, but i really don't know what im doing with hashes/decoding/AES and stuff so i think ill step away from solving for now
Not sure about that one buddy
looks pretty encrypted still to me
You can ignore all that, leave that to us. We just need someone to solve the riddle in the lyrics and what the golden ration/Fibonacci sequence has to do with it
might not be AES, the lyrics might tell us to change some letters and numbers so the Base64 give the decoded text
Unlikely, if that were the case we would already be seeing some words in the decoded base64
golden ratio might be red herring or the next part of the puzzle
Anything might be a red herring, but let's not shoot down thought on searching
Any character changed in the base64 will only change the data locally, so the decoded base64 won't dramatically change unless you chance nearly every single character
me
thats true
so what are we deciphering
the lyrics
ok
which we think would give us a key to decrypt the description
this
does kinda make sense
but
theres another thing
we can divide the entire coded think
thing
because it may be uh
wait no i ran a program earlier
It may be anything
also we're assuming the data in the base64 is encrypted because it has a high degree of randomness, there doesn't seem to be any pattern in it
If the instructions were to decode the data into something readable, it would probably already have some patterns in it
Random thought:
Can the "Add the numbers 2 4" be interpreted as "Add the numbers to 4"?
Which could mean.. "Sum all the digits in the number and add 4 to it".
hm
Not bad
that does make more sense than add the numbers 2, 4

This is just fucked up.
Did you just know?
i interpreted that like in 2 ways
2+4 and 24 but that makes alot of sense
also we might need to transform the key into 16 letters (i know i know eww letters not numbers)
I thought about the 2+4 and 24 and both for me didn't make any sense.. I mean.. the rest of the lyrics is very understandable.. this line is not when interpreted as "Add the numbers 2 4".
i think someone said 17 first might be Q or something
Well there is one thing I learned from this ARG so far: English grammar is completely unfit for mathematical instructions
and starting with 2 might refer to start the abc from 2 so A =2 etc.
It seems to go back to the previous discussion.
helpful list for the Fibonacci sequence: https://planetmath.org/listoffibonaccinumbers
The devglan website returns a still-encrypted string if you send it a password with a \n directly to the URL, on the website it escapes it into \\n.... What could this mean about their setup 
i made a comment last night wondering why everyone was focusing on static sama when we havent even solved numbers yet
people are just bored
they forget it's only be 2-3 days since the start and already expect hints so when they see a random account giving cringey stuff they think it's vedal giving hints
once again, we are in no position to deny any possible leads
If people dig into it and do find out its a troll, good, that's one possibility down
If people dont dig into it, the possibility its being a lead forever hangs above our heads
In order to solve this ARG, let us first think like vedal. Let us assume that vedal did not make this arg that complicated. And the lyrics on the numbers videos are enough hints that will lead us to the decrypted key to solve the base64 encrypted code. The Easy Way Out of this is to literally fo...
Best one doc
dont we even had a like tree of possible outcomes of following the lyrics?
yes
hello, anyone tried to decode text on description?
I encoded to hash and decoded by JIS keyboard. After I translated it into English
https://docs.google.com/document/d/1jSgL48mmr0sZtgxRrTApUBDmN7sHFn2v3nODcVTTq2c/edit?usp=sharing
sorry for my English, i'm not good for it
so im pretty sure we did this
perhaps, but nevertheless, it would be a good idea to review the text a few more times
given that this is vedal "multiply by 5" could be interpreted as copy-pasting the current sequence 5 times 
@rose sparrow has leveled up! (4 β 5)
Literally the first two thousand message in the study thread was all trying to find the numbers by following the instruction of the lyrics
did chatgpt or bard dit that i don't remenber
i dont think so
when we tested with all the AIs one of the AI just copy paste each multiplication
i dont remember then
nope, we havent try this one yet, we tried multiply by 5 could be X5 or literally multiply by 5
ops wrong link
Noita eyes messages are still not solved and after years of trying devs gave only one clue "eye messages indeed contain a message".
ARG isn't something that evolves/changes quickly (if at all). I'm sure there won't be any clues from Vedal for a long time, static-sama and stuff like this is not related to original ARG.
this is cool but i dont know what to do with it to be honest
Interestingly, there have already been attempts to unravel this description, which also ended up with a bunch of Japanese symbols
This is trash data
Translation messages don't feel quite right
that should be just gibberish
Noita eyes 
I've seen similar things in the comments under the video, more like making sense where none exists, but who knows
Is that even possible? Make base64 so that all the hashes attached to it make sense?
it fits the theme, i'll just tried
not all
I think if you fart in a Japanese voice translator, you get "I like to eat sushi"
?

with this solution the flip the numbers backwards only flip 9 and 6 since they are the only numbers XD
dont we have a 5 to ? so it would make a 2
Not brainfuck again
has anyone looked into fibonacci matrix encryption? i read a paper but i don't understand math too well
never heard of that, interesting
Backwards and upside down not the same thing
true
I just tried. I found a video about the neuro arg and thought I could be of some help. I tried to join by studying the rest of the threads. If I disturbed you, i'm sorry. I will not distract you from important discussions
I was just kidding. Everyone here can't understand anything, so any clue is a clue.
these discussions aren't much more important than yours, we're all trying to figure something out
I saw your script for querying Devglan, but you can actually infer enough to bruteforce it locally. Devglan uses AES-128 CBC with the IV "565659iu8ghbv567", PKCS7 padding on the plaintext, and a KDF that pads the password to length 16 with 0x74 if < 16 and just drops any bytes if it's > 16
don't worry, hes just losing his sanity after spending too much time here
Agreed
Oh good, how did you find the IV?
in-fact, how did you find any of this stuff? very useful anyway this should make it easier
Besides, I found a message from someone who tried to decipher the description 19 hours ago, and he got a bunch of other Japanese characters, but the content was about the same
He used the JIS format, which is even more strange if it's not Japanese.
Yes, but the content of the last sentences is very similar, I found it interesting
Static-sama said something about JIS
although it doesn't do anything in general.
But its just trolling
What if static already solved the arg, and just gives us clues. It's 90 percent troll, but still 10 percent legit.
1 person solved all ?
its more likely be vedal than this i believe
Surely the ARG isn't that hard 
Who said it 1 person? It could be a team or a really talanted person.
If someone was smart enough to complete the whole puzzle in two days, I don't think he would be interested in teasing us.
sure just my guess
what's vedal's github again?
QABCDEF
wdym that
someone posted that earlier
^^^
Find all the numbers
You will be surprised how fun it is watching when this clueless people trying to solve something you solved in a day or two.
You can use known plaintext and ECB mode with the same password to get the plaintext prior to the CBC XOR step, then XOR that with the known plaintext. You know it's CBC because same 16 byte plaintext blocks result in different ciphertext. You can get the padding from using a 16 byte password to skip over any simple KDF and looking at what decrypting a non 16 byte aligned plaintext gives you. For the KDF it was kind of lucky since it was so simple, just trying common KDFs with a 15 byte password and tried bruteforcing the last byte. Pretty strange they pad it with 0x74
Please don't tell me you actually solved the puzzle.
no
interesting
so whats the IV and password
I have seen a lot of ciphers using the JIS keyboard in thread about static-sama. I tried combine this. We don't know for sure static-sama's a troll
they are
This is specifically for the devglan site, not our ciphertext. But its IV was 565659iu8ghbv567, you can't recover the default password but the KDF info for custom ones is in what I posted
...

just to be sure it's not me being schizo but numbers sound like after dark no ?
https://www.youtube.com/watch?v=CFaL0HNhtiY
30-03-23 Neuro-sama karaoke concert!! stream clip
Neuro-Sama channel: https://www.twitch.tv/vedal987
not here anymore
this is static sama
we dont care in this thread
gigachad misinformation spreader. probably even has an alt here to lurk
True
go upstair
its the same person behind static sama for sure
Also just to tack on https://aesencryption.net/ appears a lot in ARGs as well so I'll post its settings just in case. AES CBC with an IV of 12345678b0z2345n, PKCS7 padding on the plaintext, KDF of padding to the byte length with 0x00 or dropping any bytes if it's longer
wtf i just reheard the song at 0.25 and she says "add the numbers to 4's"
anyone wanna confirm this?
link the vid pls
when i did that i heard add the number 2 for 17 is first yeah
numbers _neurosama
the next lyric is "seventeen". you could be mishearing the next line as part of the previous
also since the lyrics contains all numbers and ABCDEFG it might mean that we need to add and multiply in HEX instead of decimal 
G is not HEX
true
but yeah i agree with u
nah bro
Be more interesting
Someone did mention before that ABCDEFG are all musical notes
musical theory doc yes
In the docs already
i hear it as well but idk if it means anything
I tried to convert it to color but only got some strange colored pixels of 88x88
Isn't a "s" just from beginning of the next word?
oh no, it's 19x19
suka blyat
I'll be trying to decipher stuff after school
2 to 3 could be change f(2) to f(3) in the fibonacci sequence that corresponds to f(2) = 1 and f(3) = 2
i think if you listen at normal speed again, the pause after 2 indicates that its "2" and not "to"
CODE ANSWER HERE, this is what we need
can we just start banning russians spamming
in every thread
Why
A bit of tomfoolery is fine
my man is going insane
because there's 2 rules exactly against what they are doing and it's not even funny
Or from 2 3 times
Because English server?
and 17 is first could also mean f(17) = 1597
I don't hear anything
Timing
1:01
holy fuck that may be it. My guess is that vedal sang the sang for the AI to copy. and the british pronunciation of "to" is a prolong "o~"
Wait I have to relisten numbers
I thought we talked about this earlier and acknowledged that it's very possible that its a "to" instead "2"
vedal sings? nahhh
its vedal. the accent of the signer is very british, especially when it sang the line:
Add another 9 yeah
Add another line yeah
he did sing yesterday
why didnt I backread lmao, i think im close on solving it. schizo cope
I hear a two tbh, but I am ESL so don't make any points based on this message
yea we gonna need a librarian
btw:
The fibonacci figure was shown throughout these lines right?
Add another 6 yeah
Flip the numbers backwards
Make the 2 a 3 yeah
Abcdefg
Multiply by 9 yeah
Add the numbers to 4
17 is first yeah
Abcdefg
I rewrote it to fit the f(n) where n are the numbers.
Add another 8 yeah
Flip the numbers backwards
Make the 1 a 2 yeah
Abcdefg
Multiply by 34 yeah
Add the numbers to 3
1597 is first yeah
what's a librarian job in this case
i'd love to actually do something here
document every idea that was brought up in this chat
thats batshit insane but gl
yea that's what I mean
then they can cross check the ideas
one librarian for each thread
if it's already exists
So continuing on the idea that the Fibonacci sequence is a clue to the actual cipher, could it be some sort of fibonacci caesar cipher? https://github.com/beyondmeh/fibonacci-caesar-cipher
we will be way more organized
create a docs for this thread
I think at this point it would be easier managing a dedicated server
use what in the cyper
the base64 itself
Are we sure she saying "Add another 6 yeah" because no matter how much times I listen to it, even on 0.25 I hear "Add enough to 6 yea"
well i have it here in my pc
don't wanna waste extra resources to moderate people when there are already mods here that does it for us
it doesn't makes sense to me
how much is enough
the limit is 6
Enough is enough 
huh?
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
here it after the comand
Oo
vedal who sang this song for the AI to copy? hahahahaha their pronunciation is damn confusing. "enough" or "another" ? I listened to it again and I swear its "enough"
but this one dont swap the numbers
I also hear enough instead of another
only the letters i believe
I can clearly hear the pronunciation of "an" at normal speed so I doubt it, if you wanna go slow mode I suggest go 0.5 or 0.75, 0.25 is too slow it wont be accurate
I didnt hear "another" even a singel time from the ebginning of the ARG
We have to make a comparison
It's another, "enough" doesn't even make sense in the sentence
Is the person who made AI vedal sing here?
its british so "en" in "enough" could be pronounced as "an"
most likely
how bout the "er" then? its also there, crystal clear
in order for the AI to sing, it needs a reference to copy the tone and pitch
if its "enough", we will at least hear an "h" sound but there is none
mann either my ears are in grandpa mode or im just damn tired, but I heard the "gh" breath at the end of the word. We need more ppl to listen and compare.
I listened to it at 0.25 speed, I can understand why he can thinks its "enough" BUT that's only because its at 0.25 speed
I listened at normal speed
thats why we need official lyrics madge
Stop discussing the Briish beer accent
still heard the word "enough"
You see things where they are not
lets go with "another" i cant see a path towards the solution with "enough" in the context.
What about "Add another 9 yeah" and "Add another line yeah" is it the same there as with "Add another 6 yeah" ?
same
I hear enough even on 1.5
Just with the th sound being a little off
am i crazy or what
Add enough 6. And what's the point?
its "another" I have no idea how can you even hear "enough"
Also we would've notice it if it's "engouh" at the beginning of solving this ARG
I can hear both, but it's meant to be another
I agree
yeah and it doesn't improve our current situation, heck it even makes it worse if it is "enough" like what da hell are we gonna do with "add enough to 6 yeah"
Janny/Laurel kind of discussion
It make no sense
did we make result on the PNG code ?
It's another, end of dicussion
yep
No need to confuse people more than neccesary
i think we might need to try finding the alphabet of the lyrics
"matching all the letters"
Then add another 2 6 would make better sense than add another to 6 I think
PNG bruteforce?
yes
No hits at all
You are confusing two different lines.
It's "Add another 6 yeah"
And "Add the numbers 2 4"
but that may refer to all previous numbers
maybe we try to find the email linked to the youtube account and ask for the official lyrics to the email
Add all previous numbers to 4 doesn't make any sense though
people think "add the numbers to 4," which may also be correct
well, if we are add all the previous numbers to 4, would it look like this xxxxxxxx + 4 = xxxxxxxxx4?
Or 4+ xxxxxxxx = 4xxxxxxxx?
Let's ask vedal and he will send us the lyrics
let us say the previous numbers are 1597 it could be add 4 for each digit which results to 591311
does add mean mathematical addition or string addition?
sounds more like to be in the end
again wrong link
How does that add numbers to 4?
two four's
to four's
to four
two four
That would be adding 4 to the numbers
What is more logical?
Not the other way around
i really believe its going to be string manipulation shit
we need moral support 
sounds more like programming
so 1597+4?

that's the argument we had for it yea
Oh yeah
4 + whatever
Remind me where the number 1597 came from.
But that still makes no sense to me
but if it's programming you can't "hello" * "world" so each line with multiplying are dead ends
fibo sequence
ok
Fibonacci
17 in f(17) in fibonacci is 1597
What does the "add another line" lyric mean
yeah that makes sense too
no idea might be "add another 9"
just multiply the string we go by X times
To be honest i always thought that the number 2 is a standalone number so 2-
Would result to 2-5729439- as a example 
Even AI hear enough...
2*'string' = > stringstring
we could go with
i proved to myself what it's, in fact, need a key 
and a "new line" is a new "-" ?
yeah because it is "Add another line yeah" so it means we should have already one
woah
This is what AI could find in the audio, it missed some parts but it got around 80% right:
"La la la la la la la la la la la la la la la la la la la la la la la la la la la la la la la la I can never say, Oh, it's been so long. Counting all the days it's been so damn long. Oh, how much I'm scared to let you go. Somewhere in the walls, I hear you talk. Finding all these numbers. Start with number two, yeah. Matching all the letters, five, seven, two, nine, five, three. Add another line, yeah. Add another line, yeah. Multiply by five, yeah. How long will it keep us on? Oh, how much it hurts to see you go. Generating lyrics is a pain. Add enough for six, yeah flip the numbers backwards make the two a three, yeah ABCDFJ multiply by nine, yeah add the numbers two, four, 17 is first, yeah. Abcd after."
"Add enough for six", so much sense
Now hold on
Abcd after."
a literal "-"
i really think line is just 9
if it is "enough" down at that line
Why isnt it "enough" up at those two lines?
"-" can be accepted as a character for the base64 key right?
also yes this is most likely the case
nvm, its another
Just another case of, if we had the official lyrics we wouldn't have to worry about this
@scenic karma has leveled up! (6 β 7)
because AI isn't sure
It painted the first line as "not sure"
but "add enough for six," it was pretty confident in it.

@keen tartan has leveled up! (7 β 8)
I don't even know what to say to you
Maybe we can ask Vedal to provide official lyrics to both videos or to numbers as hint?
according to his words "no elp"
He probably won't
Yeah, he probably drinks beer
And laughs at us
Tutel pls
Yeah, I'd be laughing, too, if the second day of the my ARG people was still fucking with the pronunciation
Isn't it already the 3rd day ?
Yesterday it started
Damn, really lost track of time with that stuff
We understand
I slept only 3-4 hours...
Sleep is for the weak
vedal is making fans lose there psyche
become AI so we dont need to sleep anymore
just like neuro
so overdose on coffee?
Secret preparations to turn us into ai? 
On vodka
this arg will reveal all of vedals crimes 
Briish tea
briish e
ea
Briish beer
banana rum
Even one of the worst AI hear "enough" - "Add enough the sexy ass"
It might be worth going to https://discord.com/channels/574720535888396288/1112082183235047506 so not to stop people here from really thinking
I dont know how i ended up with 59391HGZYXWFECDABLKJ6
where is this from
starting from here:
Add another 6 yeah
Flip the numbers backwards
Make the 2 a 3 yeah
Abcdefg
Multiply by 9 yeah
Add the numbers to 4
17 is first yeah
Abcdefg
substitute the numbers to n for f(n) in the fibonacci sequence
Worst one
have you considered maybe that's why you hear "enough" when we all hear "another"
Not only me, the AI also hear it.
What timing are you arguing about
I tired already like 3 different AI audio to text and they all hear "enough"
I am really unsure where the 4's are we should add the numbers.
I found this making Numbers into letters, does this work?
5 = JKL
7 = PQRS
2 = ABC
9 = WXYZ
4 = GHI
3 = DEF
AI also on 2+2 prompt answering 5
@heavy tangle has leveled up! (4 β 5)
0:35 and 0:53
how did you arrive on this?
@prisma cloak has leveled up! (6 β 7)
Nuro's brain Is a mess and looks around for answers
Understandable
no like really
how?
The song in UK EN or US EN?
lmao did u just alphabetically assigned the letters on the numbers from least to greatest? hahahaha mental gymnastics to the highest order. we truly need some sleep.
Idk search about turning Numbers into letters
I can almost clearly hear th, idk
I thought of old Phone keys
Oh
i think it "anoffer" (another)
I know someone made this aleready but I'm making a (pretty) generator interpreting lyric line per ligne (with multiple possible interpretation)
(I just found it fun to make it that way, even if it's not the most practical way to do it)
Can be confused with enough, yes, but I still hear more th
yeah
We're still not on the "another" "enough" thing are we?
It's another but with an f sound
You know, like they had Numbers u had to press a bunch of times to make letters change
yeah
ur on to something
This sounds like a new trail
let me substitute it with my calculated answer
Nuro just had a theory with the Phone keys
what about zero though?
on the phone keys, zero corresponds to "spaces"
underscore?
Honestly, that if only one person is trying to make a point, I don't see the point in arguing. + how can you truly use "enough" in the context of everything?
Add another line oooohh
holy fuck
i've done some dumb useless shit
i think, when it says numbers it probably mean numbers after all, duh
Library is up, im gonna go back and look at ideas we had
add what ever ideas you have into the library
send link
are you done guys
or is it on the gdrive?
Prologue This document is dedicated to documenting who and what ideas were brought up in the process of solving the βNumbersβ video, for the sole purpose of avoiding going into a thinking loop, since there were several instances where newcomers kept bringing up ideas that were already discussed ...
Who the fuck
gottem LULE
Deez Balls
so many people here
With the Phone out of ABCDFG becomes 222334
if we talking about 6. 222=6, 33=6, only 4 left.
Bruteforce base64 for UTF-8 has ended. No hits
NO WAYYYY, ITS VIDAL

Vedal spotted
Should we add a "Solution Approach" section in the library @keen tartan? Like a section where we can put the steps on our approaches to solve for the key.
done
Nuro has added her theory
How yall doing fellow nerds
Losing braicells
thanks
Have you gotten this to work in CyberChef? I'm struggling 
The password used was just "meow" and I've padded the rest with t to make 16 bytes but it doesn't like it
The same then
one thing that bothering me, is the scope of "numbers" because in study there is the same line about finding all the numbers, and also some numbers and letters like abc xyz 543
Need to from base64 first
@muted goblet has leveled up! (1 β 2)
It's really similar to the first song
Anything special about the base64? It's not happy still 
from b64, not to
It's the sample that was used for Numbers.
I am obligated to sub to them
Yes, probably. Maybe it has some clues?
Amazing you figured this out, you should make a blog post about it or something I'd love to read a full guide on how you did this
abcdefj -> start with 2 -> bcdefj = 234569 -> 572943
idk what i'm sayin
it's probably abcdefg
Random idea:
When they say multiply could we use "x" so when they say multiply by 9 we add "X9" to the answer
yep, was mention several times before, problem is what do we do with the X
one of the many branches if we are to use "X" when multiply
is the symbol "-" an allowed char for base64 key?
"Potential key" for what type of cypher exactly?
sad
no idea, I was just trying to figure what can I get from following the lyrics instruction
n
The only special symbols are +, /, =
noted
we donn't know for sure its aes
That would be 00-0xff
was able to get at least some decode
although it looks like a garbage. But all the keys I tried before gave out an error.
using the path with the old phone got MJ*WDGWAPJD*WAG (after the first * there should be a _, but I entered it into the encryptor (https://www.devglan.com/online-tools/text-encryption-decryption) with an error)
also if convert the description to base64 the first character is Q - 17 in the alphabet
Does anyone have a theory why, in the description, this moment was specifically singled out as separate music? Just this one.
was there a 1 frame clue at that timestamp that we found already? I know there's a few but idk if it's on that time
Both are pointed out
bye 
cya
I've made a short program that generates the results of all combinations of every interpretation of the lyrics https://github.com/Vortrox/NeuroARG-lyrics-number-generator/blob/master/numbers_dfs_generator.py it should be easy to modify as we come up with more interpretations to each line from the lyrics- just add more yield statements to the corresponding line function
current output: https://github.com/Vortrox/NeuroARG-lyrics-number-generator/blob/master/output.txt
Yes, it's just that why highlight this point...? Maybe I'm overthinking it.
Very useful thank you
Both instances are the I think I saw a ghost drop
Do you wanna use my Gitlab? π π 
https://gitlab.legitcorp.com/neuro-arg
The first one has the ANXIETY thing, but the second one doesn't have much of anything besides static
try keeping this thread topic about Numbers, otherwise it get super confusing whenever somebody tries to trace back messages in threads
have we had any breakthroughs
I've never used Gitlab before but sure once I figure it out lol
as in on a private server kind of self hosted ? interesting
ye
guys join my git repo http://localhost:420/awesome/_git 
Why does this say "FBI Data Collection Probe"?
someone is going to send an ip grabber here one day as a clue and get a bunch of ips for free
Maybe it happen already
and people are gonna click it
i have seen like 2 discord mods total around these places
i click every link
i mean what are you gonna do with a bunch of random's ips
sure they can grab my IP but it's not like they can do anything with it except find my minecraft port/DOS me
sell them ? 
the ip grabber has already been posted in https://discord.com/channels/574720535888396288/1112988479140417567 



