#project-skyscraper
1 messages ยท Page 10 of 1
Either that, or someone here runs that account ๐
The timestamp on the image is from 3 weeks ago interestingly
Yea interesting, that's what I was asking.
HAH, well they had those images loaded and ready to go if so.
Edited on the forum to say unknown user
Oo, it's a screenshot of the post for once
its a ss of the full log
Also
and operators, i think that means us?
MB -A ....
We have a tag for the operators now, MB-A.
Well their authorization at least
MB-A, memory bloc A?
I usually do the thing not to embed the link in line. Decided against this time.
Likely, I'm guessing that's the password protected one we're working with now
The image is Fragment 1778433240, we which is may 10, at 5:14PM UTC, we got this post at about the same time https://project-skyscraper.com/2026/05/10/security-log-171453/
Oh I'm talking about the post itself, you're all good!
Which references the TR4CE image
Just checked, not directly the password but that's a good catch!
Architect went inactive on ETARC for ~30 minutes as the posts went out. They are actively lurking again.
Ah, the password is a timestamp?
Or it's a password to access a certain timestamp
Aight.
Links to a YT vudeo
Thatโs a YouTube video code.
of course its in french
Do we want the correct timestamp or the wrong one?
We had some temporal anomaly thing going on no?
Can't hurt to throw in every epoch timestamp we have.
There is a page to drop images in
oh? where?
New security log error
Isn't that just the fragment image?
I don't think we can submit anything there
yeah
If you look at the image very small/far away, it looks nearly identical to the reddit image
Click at the very top
It definitely looks like the Reddit image unscrambled btw
Could you send the direct link?
Homepage updated
Wellz usably unscrambled
looks more pixelated
and homepage looks a lot smoother
Old homepage moved to: https://project-skyscraper.com/neural-network-status/
List of posts with a sleeping log highlighted
I don't see a highlight.
I don't either
Filename of the image confirms it's related to the May 10 post.
Oh, nevermind, that was the TR4CE image...
Yo with 12 connection attempts it's a circle what
maybe we need 16
also, the password doesn't have spellcheck
I've got something!
That's just the html for what the input says and has it to not give spellcheck
disables browser spellcheck iirc
The background of the image repeats ERROR in binary (uploading image in one sec)
Not sure if you got an answer but the size there is just the visual width of the password input.
Breaks the image neatly into cells then.
Its the period 108 from my very early notes
so we treat each colour like a seperate binary segments
Ah gotcha, thanks for that catch, doubted they'd let it be that simple
See, it was useful!
Need to hurry with dinner so I can look at this stuffffff
if its a grid, i think its a 9x10 grid
The size of the individual small pixels is 1/4 the size of the bigger background ones which I find interesting
Actual name of the image file is C3q5NYF.png
Almost home!
-# โฉ Galactic Hub Project | No Man's Sky (@nmsgalactichub.bsky.social)
Operator_Auth_Level: MB-A
Replied to galactic hub
So we are definitely the operators here?
I adopt it regardless ๐
Someone at a machine take the image into the 10 columns, shift left into right to realign and combine the values
It feels so random what's been considered a "good" input to this point.
It kinda made sense with the quote page, but it's hard to even guage what we're inputting for
I know it's on the scheduled timeline, but that BSky response at least makes it feel like they want us to consider that auth request to be an inciting incident.
ok, each colour is NOT encoding its own set, i tried with cyan in one of them, got gibberish
or maybe i didn't use the right grid
I count 10x10
huh i had 12 a second ago
the ring being open
could just be the random gen
And each grid cell is 8x5 large pixels.
oops was on mobile, so the homepage also displays the restored blocs now, nice
The ERROR repeating in the background actually is different in one place in the top right where it has QSRSV in binary https://forums.atlas-65.com/t/may-30th-security-log-password/9487/21 (link since I cant upload image)
This image is 100% sus:
I'm sure it's been discussed already. Have we had @brazen prawn take a look?
@rain verge
No gifs in this channel please
has the image been cut in any way? because it looks like the left edge is actually supposed to on the right one
Oh man, the bot gobbled my GIF...
That's Galaga
I'm cooking ๐ข
Food that is
Nope, because of the binary ascii letters always start with 010
You got it! You might finally get your DQT encoding ๐
I just got home to my proper PC, got some halfway decent bourbon, pizza is on the way, wife is out of town. I'm ready!
Oh shit, well done!
Very nice
Ah I didn't realize it was gemini output at once, did you manually review it?
currently doing that
this. is this supposed to yellow shoved in there to make it green, or just green?
ok i think our dear old gemini is finally hallucinating, bro giving me stuff about fires in some section 9 or something
Have you decomposed the RGB channels? It might be easier to keep track of them if you do.
Vertical ordering GIF: https://forums.atlas-65.com/t/may-30th-security-log-password/9487/26
good sir i hath no technical means to do that ๐ฅฒ
the best i got is long ass prompts for LLMs
It almost feels like there might be some letters in the negative space if we collapse the rows down.
Oh man, no time like an active ARG to learn GIMP! https://www.gimp.org/
but what counts as the negative space? gray, black or both?
Hate the interface on GIMP, but now that they do BC7 DDS properly, it's good for NMS
photopea on browser is close enough to photoshop if you need as well
Doesn't do BC7 properly, which is what NMS is moving to
what's BC7
A particular compression scheme for DDS image files
Apparently not important for the ARG though ๐
Could we establish a rule of no AI generated takes on things unless you have manually reviewed and can confirm it. No shade but I feel like it can introduce a ton of extra noise/confusion to the discussion imo
yeah, i'm sorry for the AI thing. i am literally zoomed into the image to double check this right now
AI also sorta ruins the fun of arg's imho
I was going to crowbar my eyes out last night on the other thread.
true
Quite a nice parallel to the NMS modding experience then ๐
Or moreso the Nexus Mods experience
The upper right side of the image breaks the background pattern.
it ruins the fun of a lot of things not that this is the time for that discussion ๐. But in terms of ARG if you wanna use it for yourself thats your own discretion, just if AI hallucination pollutes the discussion things then its just kinda destructive
Yup, posted above, spells QSRSV
I've done it, too, no worries ๐
Basically
Doesn't mean anything using the school code either
Neither that or the school code version were the password
does it look french?
It looks like if I layed on my keyboard
ah, it looks german
(so not french ๐ )
What are we looking at?
Eating now then time to rip
problem with decomposing is that the image has a nice little tint of blue everywhere, its showing up even in places where red is. i need to learn GIMP a bit more to properly figure this out
@harsh vapor have you done a (non-background) to bitplane and seen if it alters or XORs into the background plane
counter-filter until red is #FF0000?
i shalt attempt
I have not no, it's a bit difficult to isolate the background just itself since it's not very cleanly separable and the tint in the blue channel makes it kinda annoying (unless I am vastly misunderstanding what you are saying ๐ )
Searching I don't see it mentioned anywhere but a small post from 2017 about WT, but fyi there's this site that's really useful in ARG's https://www.dcode.fr/cipher-identifier
mb if y'all already know about it lol
So we're supposed to add a new memory in?
seems like it
the clip linked is kinda interesting narratively
it talks about HALs memory being wiped by a program, and here we are trying to recover lost memory blocs
Whatโs interesting is this image is timestamped to the same date as the tr4ce image/log which does not seem like a coincidence at all
adding a memory is odd though. is it a false memory, or one we recover ourselves?
The password page is for timestamp 094317. The latest sec-log is 094316
Thought also before I hit the keyboard
Flatten to 3 color range
Bit flip with high(2)
Or XOR
That's Jan 2, 1970, right after the first memory
i think the process would be what we did for the poem post, but instead of tweeting our answers, we directly send them in, wherever we have to
And: [30/5/2026 09:43:16 AM
That QSRSV cell also breaks the cell to its right the 'O' is missing a bit.
94316 is important, I think
They specifically mention TEMPORAL memory.
Maybe one of the memories that was out of order, we have to give the correct timestamps?
94316 on a google search gets a timing belt part number. I doubt that's intentional, though ๐
I think itโd be worth just feeding in all the timestamps we have
I think we are going to need the password before we can add the memory
Agreed
We might need the password to get the timestamp for the memory we're supposed to correct in the first place going by the title?
Could be
ok at my machine now, what HAS been tried so I'm not writing useless tests
its weird that the fragment timestamp is for TR4CE though, when the reddit sec log is a few days later
We've got the background bits thing, but what masks have been tried?
Realistically, nothing substantial has been tried beyond decoding the background binary.
why do i feel this image will change soon enough
i think each colour is enocding a seperate phrase
A couple of days ago ๐
Ah
like, a different one for cyan, for blue, for magenta
Iโve just done a very superficial look at the individual channels
Also, that memory timestamp is a sunday morning
there are so many links between things atp that the evidence all looks like a conspiracy board ๐
FYI, August 9th 2026 is a Sunday
Small run through brings up nothing with the random letters
We need different colors of yarn now
oh interesting...
The architect's background image on the image site was uploaded on <t:1778307123>, so the account's probablyy been active since then
Quick note to anyone in public, do NOT click anywhere else on that site. I got flashed trying to confirm that the background image names have timestamps corresponding to upload dates ๐ฅฒ
can you tell me how the OR mask is being applied here? i know the terminology but can't figure out how its applied to an image
Building the 7 color bitmasking tests now.
Seems like a fragment of something bigger (perhaps 'les')? Perhaps destabilisation will cause more to be revealed?
It's not really a mask, it's literally taking the values of each individual pixel of the image and conducting an OR operation for each of the RGB values.
ah
OR/XOR etc are the operations but the bitmask is the pattern used for the bit fiddling in the op (for or you use the bits to mask for ON position), it gets more complicated from there
All basic ops with them in reduced produce garbage too
๐ค hm
Is the forums progressing more/equal to here, or best not looked at for now?
No progression
Trying to give it a break for now ๐
The red channel is really odd to me
Looking at only the contrast patterns, it looks like only one block deviates from the background pattern, right 8, down 1
Everything else is this
The one to the right of it is different too.
block I mention is this
But only in the 'O' position (4 rows down from the top)
yeah could be, the fact that there are a few value based near collapses on values is interesting from a bitplane perspective
I don't see it, could you screenshot it?
Quicksilver SRV :trollface:
these two I believe. the one on the right of the one you sent is interesting bcz its missing a single value
oh, in the vertical banding, I see it now
Yeah, it's almost as if the errant cell overwrote that pixel.
but guys, the green bits. are they actually green, or yellow bits under the blue making it green?
that was my first thought too, does the overriden binary output readible ASCII for that character now btw?
(to save time for some)
green bits are actually green
Technically, I think 00001111 is 'Shift' in ASCII
yellow is R + G with RGB
Sweet!
I don't think we're mixing colors here. Or at the very least the background ERROR bits aren't interacting with the foreground.
oh, that's super intentional then. Reading it as the QSRSV block "shifting" data in the adjacent block may be right
Wait, do the foreground bits overlap at all with the ERROR message?
Maybe that one block with the discrepancy in the background is what's causing the error, and we have to figure out how to correct it?
Given it is one bit off from being the correct O, I'm inclined to say it was an oversight when creating the image
I'll trust you two ๐
QSRSV is definitely intentional. It's almost reminding me of the /code page though. Although it doesn't work for a simple substitution and we only have 5 letters.
As in treat B1 G7 and B1 G8 as shared regions?
My gut says that discrepancy is intentional
They're almost completely clear of colored pixels except for that one blue one. It's one of the cleanest areas in the image, save for the cell in the top left.
the green here looks suspiciously like morse code (not nearly structured enough tho)
vertically?
like, the morse is encoded vertically
wait i'm confusing with cyan
sorry
How dare you!
R I E E ?? E I I T E, yea no
I am still thinking these are masked in a way to get us the fiddled output
Gravitino balls to the wall I am getting nothing useful out of normal ops with them though
Like it feels like there is definitely some pattern happening on the green/blue channels
Maybe there is a mixer we are missing?
Hmm maight work? Not sure if that'd give a working output though
What other hints does it snake out to
One would think, especially given the colors we are working on
๐ Homepage changed (structure/text, ignoring the live counter) (+['SECURITY']) (-['System Filtering Platform'])
https://project-skyscraper.com/
I'm convinced that the image will give us the password, but on the off chance that it's for a seperate puzzle, I made a tracking sheet for password attempts so far:
How can this connect back to trace is what I keep coming back to
Wait, what is this text at the bottom?
v=xPG-VM__mwU
youtube video tag
The YT video
Slightly late - I just woke up
hmmm..
Ahh, thanks
Isnt this from an hour ago? Or am I blind?
It is
Ah gotcha
Produces junk under most ops too this way
Maybe we need to XOR the ERROR and QSRSV binary pattern and only look at colored pixels that fall either inside or outside that pattern?
Gotcha, didn't think it'd go far but I'm glad you thought it was worth checking in the first place ๐
Any reason why XOR-ing and OR-ing is something y'all've been doing?
Math / Cryptography
Oh! wait a tick (I am wrong for sure) - DQT had the stepping and shuffled 1/2/3
With the reduction to 3, maybe that gives the base + mask
Geknip theory lmao
I feel like I'm tripping spaceballs half the time I hear ya describe your processes anyway haha ๐
Makes me more glad we have ya working on it with us haha
XOR to ABBCB but looks susiciously letter-like
But, that white space may be the acceptable pattern space for the colored pixel data.
Looking over video, haven't seen the movie in forever. Things that stand out to me: Discovery, launch window in 31 days, HAL thinks it is functioning perfectly, 28 months, 9000 series, virus that destroys desired memories, monolith, Bowman and Poole = Architect and Ghost?, 34223, 30 Jan 2001, Mobius
No go, nor for TR4CE
NOR on TR4CE? :trollface:
Writing a really funky "3D stack" idea to test
Had me thinking about the original tower quote
Which had me thiking about panopticon, then the hat problem
One parity bit who can see the other, etc etc
I just tried all ~40 UTC timestamps we have as passwords and all came back as invalid
Took one for the team ๐ซก
I'll document that on the sheet so no one else has to go through the annoyance
I read that as you came back as an invalid
" ! " is missing in the chart
Wait!
That looks like the block in the image I saw
one sec
on the when we were 17 image
Tenuous
Yeah probably. Just strongly reminded me of it
Interestingly, the latest security log has WP ids 288/291 (vs like 1700/1800), these ones have been cooking for a while...
Hmm, the background 1s are never obstructed by the colored noise
what is up with the weird #151515 dots scattered throughout?
@brazen prawn is your arg word list on the repo? Heading out for a bit so I can't work the image, thought about running through our candidate passwords.
I think it is in there somewhere
Missing a lot of the newer words
Maybe it's something we ought to crowdsource for future use though to have a central wordlist
I am holding off pointing a cluster at the page like a dog with a steak infront of it
Like there are a small number of gray dots of all the same color that only show up on the darker gray background lines
Yeah
They only showed up next to the colored pixels, so I thought it might be some aliasing, but it doesn't appear on the other background color
They're in that upload before
Ah ok
They're curious
3D stack parity holder idea was fundamentally flawed and produced jack all - same issue, data is too sparce even if the parity holder passes through or flips
So simliar
Probably an unusable version of the image
Maybe sent as a teaser or hint to keep shadowcheck in mind while solving this
Ooh, flipped vertically?
They have diffrent orygial orientation
Oh interesting
What if take 2nd pic and stack them
Would the read order for the bits change if the image was read in shadowcheck's orientation?
Meaningfully that is
I just woke up and now there's a password to enter??? Have you guys figured it out yet?
Shadowcheck matches the current image's orientation, I just checked.
Yeah.
Ah, just misinformation then
already try the RGB method with that img?
Hmm, architect seems to be checking in on ETARC every 15 minutes or so. Lingering for a bit, but not actively lurking like previously.
What's the grid size again? Accounting for any cutoffs?
10x10 cells of 8x5 large pixels.
And the coloured pixels are are 1/4th the original pixel size correct? Or do they vary?
Correct, the colored pixels a 1/4 the size of the ERROR pixels.
Worth noting it lines up with the 10x10 bitplane voids in the CDAD image as well as the 50 potential data segments (no, I don't think it's part of this puzzle)
Can da vinci or golden ratio have something to do with this ?
unsure if this is helpful, but here is the image without background/properly saturated/brightened, makes the individual color channels much cleaner
Tried running it through but still garbo
Gonna try and extract all the bitstrings from those and see if anything comes up
Could the different colour channels act as a check for each other? Blue seems to have individual rows where a pixel is filled if the one to its right is filled, which is not present in the green layer.
Per colour or in the combination?
Starting with per color
Ahh ty, that is pretty much what I meant
Where is the original of this image?
Spinning up another box, need more windows lmao
https://project-skyscraper.com/sec-log-094316-full/ On this page, the "ERROR" link
Actually my plan was to isolate into the smaller subpixel grid instead of still working with the 10x10 grid
Maybe yeah
thats interesting..
Yeah Sounds like actual melody of something.
how did you make that?
I hear notes, but I wouldn't call it a melody
Yeah
More like just things that are significant
It's essentially a cleaner version of the "burst" sounds in the CDAD image
https://nsspot.herokuapp.com/imagetoaudio/
With online tool
I think it's the auditory version of you can see that there is some pattern on the image
Alright yeah, analyzing the individual bitstrings extracted from the image didn't really seem fruitful
Yeah, they produced garbage in all my runs as well
I am looking into this as freq band bursts atm
Timing for "on" runs are always multiples of 10 pixels (10 = 6291x, 20 = 2822x, 30 = 1558x, 40 = 1000x). quantized like a baud rate
Seems convincing, but so far producing garbage also
mYfReAkInEaRs
Keep in mind the french quote about simplicity.
I am simple though 
I think the closest to anything out of the noise has been Vau1t's gifs. There is certainly a pattern in the adjacent blocks, the shapes seem to morph "frame-by-frame"
That you are. I meant the decryption.
The GIFs almost remind me of the Bad Apple video that kind of became a meme.
That would be gold
The GIFs kind of feel like we're missing data. Like there's more color or pixel data that we just don't have.
I am losing faith too, we may just have the keyhole at present
QSRSV is the closest we have to anything atm
The shadowcheck is making me think: Blue ghosts (is cloned across the image). Green has gaps(?). Red acts as some sort of control.
One image, fragmented. Together they are whole (or can be made so).
I think 94316 is important as well
It may be
ERROR
-----
QSRSV
No wait, that's 3 Rs
Don't forget about ERVOR ๐
Yeah, it doesn't make a lot of sense, but it could be a vignere key. We just don't have any obvious ciphertext.
I am of the opinion that this solvable now though
The repeating "ERRORERROR..." is vigenere-esque
true
Is it possible there's something more to be gained from TR4CE? That's where this image points to via the timestamp.
I do think so, just not sure what
Is there any record of the May 10 log before it was updated?
Probably difficult to find, we didn't really pay much attention to the ARG before the system unbanning on ETARC
Some leaders on the forums might have an idea though
The more I look into the noise I'm thinking it's less and less important
Gonna bruteforce against redacted bits in a few chunks as if error is itself the redaction and go for a shower while it runs
It still looks like an NMS alphabet
Output was sadly what I figured, bit space is still too sparse, so 1070 odd lines of garbage string permutes
What happens if you lay each "layer" of the image over the other?
Is there a clear break in each row?
I suggested layering pictures before.
Anything yielded?
Many bitwise operations on many combinations, orientations, colour spaces and bit planes - so far nada
Operator Crew status request 
Other than just being geometric pixel patterns, I'm not seeing a similarity
[compiling new wordlist]
Gonna be out for a bit still, but i'll use your new wordlist and throw it at the POST form automatically once back, then clean up the password attempts spreadsheet.
have been trying a few manually rn though
I already got burned trying it haha, ensure your POST times are ~>5, I flew too close to the sun with usual ~5
Will chuck onto repo in a tick with some other stuff like the color splits etc
Haha, architect's probably got a smaller limit for ya, looking at that history ๐
We kind of got stuck on the 20x20px blocks per bit idea - but we didn't think about them being a nibble block of 10x10px bits
Not yet
goddemn
I interpreted it as such and didn't get anywhere
I was going to ping you before I started but didn't want to disturb
Yeah so far useless also
All channels, all combinations of mixing RGB with bitwise AND,OR,XOR, interpreting it row by row, and column by column
Gives a few more possible "ERROR" abstractions, but that's about it (and no useful decodes)
Come get in on the fun! (of staring at strange pixels for 4 hours straight)
I think I just stumbled onto some random babble that gave me a clue.
- In 1912, Carl Jung published the theory of The Shadow (an unconscious aspect of the personality that does not correspond with the ego ideal).
- In the 2010 clip on YouTube, the same idea is covered when talking about HAL 9000 (it had to lie)
We had the clue of "shadowcheck-1912" wrt the reddit post, which appears it could be a distorted version of this image - perhaps a literal version of shadows is something we could consider for this image (remember the tesseract thing?).
we already connected this with memory fragment 1778433240 right?
edit: yes we did
Like it's possible that the binary digests include something more (like a file or some other binary data) but that's going a step deeper than I am willing to investigate if all of them start with a ton of 0 bytes
Just seems very unlikely
i would like to bring up the quote again, simplicity is the best form of sophistication
us diving into it too much could just be hiding the answer from us
which seems to be a running gag with the architect
Yeah, it maps out to a hell of a lot of dead space in the same way - the image has an overwhelming data sparseness issue
@rose dust I attempted all variations of ROYAL, SAGE, WEAVER (caps, proper nouns, individual, all together)
๐ฆ New Bluesky post (2026-05-30T09:31:39.467Z): โOperator_Auth_Level: MB-Aโ
https://bsky.app/profile/skyscraper-prj.bsky.social
So far the password page is linked to shadow check but the image, TR4CE by the fragment timestamp, potentially atlas passes by the BSky response to the galactic hub and the HAL clip from 2010 so far right? Any other connections?
I hope we get some kind of operator merch at the end of this. I am infatuated with this ARGs theme so far.
we need a figma page for this
We could always make our own if whomever this is doesn't make any
Don't give me more projects! I'm already drowning lmao
Isn't this outdated?
what's a shadow check
I don't think we know for sure. There was a shadow ban check that showed a second post on Reddit by Architect
The sec log about the Reddit post
ah
I think Pixy found that. But I don't think 1912 connected to it
Without negative timestamps, it's going to be hard to connect 1912 to anything, though
Oh btw Iโm working on an auto cypher - trying to keep up with the latest things but youโll be able to input a coded cypher and itโll pop out the correct layout (used through the bot hopefully)
I'll add that to the sheet!
I'm going to repeat that I think it's about psychology, and how it relates to AIs going insane due to unresolvable conflicts
reminds me of rampancy from halo
My intuition says it wouldn't use those three, but it doesn't hurt to add it to the list. I think it's fan-made, commissioned but not perfectly coordinated, or they are avoiding really obvious connects to official HG stuff
Oh I'm keeping a list of invalid attempts so we don't work too hard on guessing haha
So these failing is good groundwork for others to follow
what all passwords have been tried so far?
https://docs.google.com/spreadsheets/d/1_3dKZrJ7IDcUWqYDTHTWJjS0hFwNEKpDxcItedFde24/edit?gid=0#gid=0
I'm also curious about the whole "French" thing. I think if a Frenchman were trying to make an ARG for NMS, they would consciously realize most players are native English speakers and make it English. It being in French seems something intentional and significant to me. Wasn't the WT corporation French, or something like that?
@rose dust it's in the repo root and yes some of it is wacko
- have been mixing it with the usual suspects (https://github.com/kkrypt0nn/wordlists/tree/main/wordlists)
The split colors, gifs, reports, dumps, and blah blah are also there for the new image
Thanks! I'll hook up my requests with this as soon as I'm back!
If nothing else, it'll knock off a good chunk of attempts hopefully ๐
The password is probably passed through the cipher
Since they just recently gave us the cipher, yeah, I think it's connected
should we wait for any communication in the inbox then?
Oh absolutely. I'm sure the french bit is in character, and not precisely (though could be) indicative of region
do we have any indication that the password is 20 characters?
or was that some bit of code that doesn't actually confirm it?
Once I am done with this instance I'm gonna kill it and roll over the IP and join in again as well (will go bottom up + common wordlists mixers)
are we using this to brute force it?
Mostly just using it to knock most likely candidates out. I doubt it'll be any one of these
Same
is it trying the same in french?
It's just the visual char limit for the input field
Hi there. Was randomly looking at solutions regarding "MB-A". I found a reference to Azure -used by HG- and MB-A might be something like Memory Bandwith Allocation. Don't have any network admin knowledge so just throwing that under the bus
I know that in some cases you add suffix to cmd at the end to give authorization to a user
I do think that MB-A might have a clue in it. Memory Bandwidth Allocation doesn't seem quite right, but a good track. Would need to be a two word MB, followed by a one word A subset/specific (I think, personally)
I've just tried (unsuccesfully) admin
Memory Bloc - Addition? As in, our access level is restricted towards only adding memories at this point (if we can get in)
That makes sense, yeah
Yeah I think that tracks
Well I woke up to a lot happening
I just realized that the particle page is no longer there...๐คฃ
(but not really)
Does anyone know what these dates mean when you go to the main page of project skyscraper?
They're log entries
Interesting
feel free to peruse all known pages if you're new: https://project-skyscraper.com/sitemap-1.xml
we have a lot of documentation on our investigation here too: https://forums.atlas-65.com/c/project-skyscraper/54
Did someone scan the comments section of the YT video already ? I've tried as well but no luck till yet...
in this clip the scientist mentions "mobius strip" and now all i can think is...
ITS MORBIN' TIME
๐
Gave it a shot, but a lazy one
that's where I'm at tbh ๐
It's already been tried as a password funnily enough ๐
I still think code could be time stamp numbers
All of the memory bloc timestamps from the site have been tried already
None of those ones worked
Maybe that ine is not on the site
Yt vid might be the hint about the dates
The ones mentioned in the clip were tried too
Thought directly from the french subtitles
Doubt they'd want us to guess 24 hours worth of timestamps for several dates though
I said . Hint not the direct date.
I really hope they are logging the attempts for a smile later
I tried 94316 because I think it's important, but that didn't work
Yea that's good, I'm just mentioning that the actual dates were tried already, in case anyone was planning on going through those again ๐
Oo wait I saw this mentioned before too, but why do we think it's important? I think I missed the train if thought for that
I also tried 093416, 94317, and 094317, fyi
It's in two or three places
Has anyone tried the sleep mode excel numbers yet?
Is it the one the latest log?
https://project-skyscraper.com/request-memory-timestamp-094317/
And it's Jan 2 1970, just after the first memory
I think there's another reference that I forgot right now
Ah treating it as a unix timestamp?
Yeah, as a timestamp it's that
Not conclusive in any way, just that it's in a few places that were just posted yesterday, so probably important
Tried imput in reverse ?
All good! I think it's worth keeping in mind, I was just confused about the date bcz I haven't seen such small Unix timestamps in a while ๐
you're the goat for this
doing a deep dive on this image due to patterns
glad to see people still at work on this
ARG wordlist is no go (no surprise lol)
Havant seen that tried yet. You could give it a go if you'd like
feels like a game of pixel sudoku or something lmao
I also temporarily forgot I needed to log out for the sheet 
Oh nice(Maybe not in the best way)!
At least this lets us ignore most of the other ARG clues so far
For sure
Saw someone convert it to a gif too, if nothing else, it looked really cool haha
that was sick tbh
There's a few runs in the repo for the gif (diff, clean, with/without resolve, etc)
Currently looking at some split frame overlap decodes but not hopeful
Am I right that only one of the segments has a different-shaped background?
They all seem to have the first pattern except one empty cell which has the second
wellp.
I divided the image evenly into 160x160 squares (the bottom remaining 40px left out), then combined them into one, and nothing really sticks out
Ah, yeah, I was seeing the QSRSV one, but I see the ERR.R one next to it now
I have variations and still not seeing it
but I want to highlight that each little "pixel" is 10px wide, so these squares are 16x16
not 16px by 16px, but 16 squares by 16 squares
20ร20 dot R channel LSB decode always resolves to ERROR so all you get is a shifted ERROR output at 16x16 like: ERRORERR ERRORERR unless I am misunderstanding what you mean
I'm just saying each colored square is 10px by 10px
so in the 160px by 160px images I just shared, there are 16 by 16 squares
I have tried major events timestamps that happened in the movies 2001 & 2010 as utc format using gemini .nothing worked
Event Date and UTC Unix Timestamp :
First Monolith found on the Moon 1999 915148800,
Discovery Mission to Jupiter 2001 978307200,
HAL 9000 malfunction and deactivation 2001 978307200,
Dave Bowman enters the Monolith 2001 978307200,
Leonov Mission to Jupiter 2010 1262304000,
HAL 9000 reactivated 2010 1262304000,
Discovery and Leonov leave Jupiter 2010 1262304000,
Jupiter becomes a second sun (Lucifer) 2010 1262304000
Decanting the 2010 script into a wordlist ๐
2001, not 2010?
@turbid eagle
No gifs in this channel please
2010 yeah, edited
Oh well, the password is not kcmaufc (neuf ans, 9 years, run through the cipher)
We still donโt know it right?
correct
Well, hopefully we get the key drop soon so we can stop doing loopy stuff lol, gonna get some sleep
Trying to catch up. HAL was turned off due to malfunction. When it was turned on again, a time worm was used to selectively delete the files that caused the malfunction, correct? Sounds like what is happening to the System..
True
Sorry ๐ How did you get "ERROR" from the image?
the first things i thought of were:
conway's game of life
the NMS alphabet
the Arecibo message
and those videos where someone puts like mspaint through audio and video converters
We were looking at the 'background' of the image, the light grey / dark grey (20x20px) and encoding those blocks to bits in a 8x5 block region
oh the website changed
Okay, thanks, that's interesting ๐

Some quick and dirty marked up explainer images (one is ERROR, the other is QSRSV):
also just to be clear i'm absolutely terrible at ARG's lol
Okay, got it! So logically, we should be looking on the "QSRSV" side
Don't worry, I've never done one before!
I'll leave the others to explain further, I've gotta get some sleep
thanks ๐ Good night !
Does anyone have the SM pics of "When the sky stops being the limit and becomes the threat"
i'm looking for it on the forum and cannot find it
nvm i found them
each square being treated as a binary 1 or 0 (in rows of 8 digits) result in binary to text conversion: ERROR
try it out: 01000101 01010010 01010010 01001111 01010010
I'm going through every iteration to see if there are any other words
(there aren't) lol
01000101 01010010 01010010 01001111 01010010 ERROR
01010001 01010011 01010010 01010011 01010110 QSRSV
01000101 01010010 01010010 00001111 01010010 ERRR
its not confirmed to be Sean btw
That account WAS mentioned by Sean to be his back in 2020, but ever since 2025 they've had a new instagram page which is linked everywhere
i see
And now the old account (where the posts come from) is in a weird status of not being confirmed as either still run by Sean or not.
Would it hurt anything to try and ask HG
Since nothing official from HG points to it anymore
I dmed the account about it but I highly doubt we'll get a response. You could try contacting HG (maybe try reaching out to Tim Woodley directly?) though be prepared to be ghosted harder than by the architect.
I guess you could frame the question like "I think someone's impersonating y'all" to try n get a response but that's still an iffy thing to pull
do you have a link to the IG account? when I email them I'll link to it for reference
It wouldn't hurt to ask, but don't expect a response. They don't even answer when people knock on their physical door bearing gifts
i do the same thing ordering a pizza tbh
i leave instructions to leave it on a table near the door
That's what we do, but it's more common in Korea
thanks bomber
Have you looked at the filename of the pixel image? It seems very password-like. C3q5NYF
??
here in the US I dont think its common practice at all, I feel like a raccoon waiting for the delivery guy to leave before i grab it
yep, it didnt work
I couldn't find any translation of that image suffix. Maybe try it without any transformation as the password?
Oh damn
That seems too easy, but worth trying
I just found out that HG "poached" Tim Woodley from 505 after getting them to publish the Xbox port
No wonder i havent heard of them in a while in relation to NMS
I'm not aware of the context. I feel like a minor nerd now
I tried C3q5NYF itself, yes, but it could be base64 code or something else
this filename is strange
I figure whatever the password is will look like something significant and recognizable. I tried (and I'm sure many others did) to convert that into something and couldn't find anything
505 Games used to handle publishing for NMS on Xbox iirc, at least on launch. They've been self publishing on other platforms since, but it turns out that Tim Woodley (their current head of publishing) used to work at 505 Games immediately before joining HG, so he likely worked with them for the Xbox port too?
Source: https://www.linkedin.com/posts/tim-woodley-0870242_publishing-developer-activity-6518735806666985472-TvLZ
After almost 11 amazing years at 505 Games, today is my first day at Hello Games as their first Head of Publishing. Seriously impressive what impact this small hard-working dev team have achieved without any dedicated publishing personnel. Really exciting to think what we can achieve with the brakes taken off. Canโt wait to get stuck in, be de...
completely not relevant to the channel though my bad ๐
But still cool, particularly with their breadth of publishing and also publishing for at least one other indie. Thanks for the details
The Architect is lurking on ETARC again rn
"what's taking them so long"
well i sent the email to HG
we'll see if they respond
probably wont be until tomorrow sometime since it's after-hours in the UK now
plus a weekend
Probably won't be ever ๐
oh yeah i forgot
yeah that's possible too
I sent one in too just in case. Might make the "impersonation" seem more concerning?
but it only took like five minutes to draft and send
in return for a possible breakthrough of how we see this ARG
Which, at the rate of misinformation, it definitely could be if people start taking a potential fan account as HG's recognition of the ARG
I didn't mention the ARG or anything when I sent the email
oh same, just that there's a lil confusion over accounts and / or a possible imposter
What impersonation is going on? Maybe I missed something
It's not so much impersonation as it is confusion around who's controlling the thing
and that misconception Could theoretically be seen as impersonation
if people start thinking the acc is official when/if it's not
there's an intstagram account (linked above), which used to be run by Sean (linked in 2020) which posted the ARG's tagline
If I were them, confusion is fine and completely different from impersonation
yeah
Gotcha thanks
idk how you'd even hand-off an IG account so my personal thoughts are that Sean still runs it
The issue is, ever since they launched their latest bsky / instagram accounts, we have no clue if the old one is still run by HG or not since they scrubbed any links to it.
but maybe they made an arrangement sometime idk
They still post trailers and such on that account and have a mostly "official" looking bio, and actually have more followers than the real account, so if it is fan made its definitely superseding the real one in people's feeds
I think they posted two images of the expedition with the captions When the sky stops being the limit and and becomes a threat respectively
theres a screenshot on ETARC by Sheral
Thanks, 4am for me so I'm not going to dig too much ๐
haha, glad im not the only one going through this ARG sleepless
๐ค
Okay, thanks 
Wait thereโs a new arg?
yes, welcome! unconfirmed if official
Sweet whatโs the basics so far?
I'd start here: https://forums.atlas-65.com/t/project-skyscraper-welcome-info/9206
But it's called Project Skyscraper, and something is waking up and missing its memories which will likely need to be restored. We're currently in the process of trying to find a password
First of all, a warm welcome to our forum and โnormallyโ quiet little place to discuss anything No Manโs Sky. We have received a lot of traffic ever since this likely fan made ARG popped up and literally appeared to โkickโ off on here. In hopes to prevent any issues and provide some quick info to get you started, please read up on som...
Sweet canโt wait til we get bob back
lol saaaaame, bob is the goat
I don't wanna know the amount of time I've spent staring at these pixels today ๐
hey did y'all notice the subtitles of this 2010 clip are in French?
I'm slow, so forgive me
The thing that is sticking out to me after taking a decent break is, why is there such a clean break in the noise at the bottom, then nearly every bottom pixel is covered
There seems to be some patternish with the rest of the nosie, that just is really sticking out to me for some reason
Yes, the characters "The Architect" and "Ghost" have communicated in French (https://project-skyscraper.com/inbox/) and the timezone of the site is GMT+1
right which is why it stood out to me that the 2010 clip also contains French subtitles
possible suggestion of a French password
Yeah the French angle is being played up quite a bit
Guys I think the architectโฆ. he may be French ๐จ
A Frenchman\woman wouldn't be working until Monday at 10am
๐ And that's when we're not on strike.
Oh no
Running through all caesar ciphers of QSRSV, with shift amount 22 you get MONOR. It's the closest thing I've gotten to anything, it's the name of some town in Budapest (and the name of some French travel vlogger ๐)
The issue with having this little text is it can be very easily be massaged into looking like some real text with ciphers
I'm not allowed to share pics, but the glitched image is Fragment 1778433240, and wouldn't it belong in memory bloc 1778433240? https://project-skyscraper.com/2026/05/10/sec-log-171452/
I think we get closer to the password by using our current "Status: Memory bloc addition authorized" to get the image in the correct memory bloc and maybe it will update or progress?
That's an interesting point, I haven't thought about it in that way specifically. I think if there is anything major we are missing with how we've been looking at it so far is the connection to the previous log/TRACE security log
Hmm, could we try adding the new pic to the reddit one? ๐ค
I'm pretty sure that started showing up today
It was on one of the previous security logs on the footer, but now it shows as the footer on the main page
It's always been on some pages, but I don't recall if it was on the index
Did anyone figure out anything about the image in the Jan 1970 entry?
The skyscraper pic? We know it's the Atlas tower, but not if there's any meaning to the modifications
We don't really know, yet
question, how were the letters QSRSV taken from the fragment image, I mean this one
it was a process.
Step 1 was to adjust the color levels so that the blue, cyan, magenta, green were basically gone, then enhance the contrast of the darker background layer
Step 2 was to divide it evenly into a 10 by 10 grid of 160x100px squares.
Each of those squares contained 8-column, 5-row grids of light/dark squares.
Step 3 was converting those light/dark squares into binary (light=0, dark=1)
Step 4 was converting binary to ASCII text, which, for 98 of the 100 squares, resulted in "ERROR" (literally, that was the binary translation), and for two of the squares on row 2, columns 8 and 9, "QSRSV" and "ERRR"
the result was this
white = 0
black = 1
repeating grid of the binary message, "ERROR"
ah interesting
@brazen prawn did the grunt work while we watched in awe
Always glad to be in the presence of people much smarter than me
Where does this leave us for the day, then? We'll see some activity about 8 hours from now, if the last couple days are anything to go off of
im just working through the stuff myself, seeing if there's anything that hasn't been considered
So, I'm pretty sure this image is just a blurred out version of the fragment image but before I knew about that, I thought it might be one of those optical illusion type of images where you can cross your eye to see a hidden image
probably tried already idk
It's either the new image, blurred (like you said) or it's a less corrupted version of whatever we're trying to restore. Maybe both lol
there sure seems to be a lot more green in this one than there is in the new pixellated version
yeah the magenta is also in different spots
Pre coffee thought but what if the QSRSV blocks "loss" bit fiddle is actually the important operation and needs to be marched?
I want to help
can you elaborate on what marching entails?
I'm also puzzled by the "ERRR" block where "" is ASCII 15, the non-print "shift-in" command
context:
It was historically used in teletype systems to signal a switch from the upper character set (often defined by a hardware switch or paper tape) to the lower character set.
I guess it's from a 7-bit ASCII character set
in Unicode if it displayed properly, we'd see, "ERRโOR"
Gonna have my coffee and breakie and clunk away at it - but what I mean is maybe if we look at the word block as a set and the "bit flips" as result of a movement operation not a flip, the movement patterns can be used to attempt to march/move other main blocks or sub blocks (bit or nibble). Whatever the answer is it needs to be a significant varied shift to move away from the low entropy read
I think I understand (I probably don't), I'll try to take a look
Either way I'm stoked you're back on!
The problem is in isolation it can't be a simple movement, the window has to open to include neighbouring blocks etc.
I'm currently thinking it might be a purposefully-mangled set of glyphs from the Hyper Light Drifter alphabet. HLD was one of the games in the book from the /Code entry, the artistic style matches, and it has monoliths in common with 2010. We also had "DR1FTING" in the CSS.
well we do have two blocks, so we get to witness the movement twice if there is one, which could give clues into how to continue it forward
The problem is stuff like this seems way too over-engineered and simple ops like internal block op -> blocks as an op doesn't change the entropy enough
Maybe it was truly a pre-coffee thought ๐
Gonna write some tests anyway
Are the HLD glyphs all 4x3?
Ah, bummer
I could be on the wrong track, but I feel like we're missing a clue, and there must be something in the pile of stuff we have already. I might have leaped to the wrong one, but I did go through every game in the book.
Probably important to keep in mind that the Da Vinci quote was about simplicity.
Also kinda my point yeah, we're entering severe over-engineered territory and it needs to be kind of approachable by the majority of people
Approachable but not obvious, yeah
Maybe a rewatch of 2010 is on the books for the day
I think we're still in a phase where it might even be somewhat obvious.
Have we bruteforced the password input with NMS-themed wordlists? (Not advocating for this but it's an idea)
These are 3 high by 4 wide, right?
yeah 4 x 3
A little yeah
I've momentarily abandoned the image in favor of 2010
Taking a step back and a new angle: 7 bit punched tape
I was hoping you'd look into that
Here's an odd idea: do we have to give the timestamps for future memory blocks/past memory blocks from before we started seeing them using the standard cadence we've identified?
But also in what format do we have to give memory blocks/timestamps anyway?
Or the accurate timestamps for the last ones which were not accurate?
I'm in the process of actually reviewing all sleep/sec-log timestamps.
Some of them are literally hundreds of years off
I will say an ARG i worked on a while ago almost got closed because people were trying to brute force a password (too much on the server or something)
Can also ruin the fun
I can't see a significant read path via the tape idea either with the 'base' layers, but that is to be expected. You just get the same shifted pattern data (obviously) but the outputs are all nonsense. Same with the 'content' layers masking the 'base' layers, the output is just stuff like Band 0: P""s" p""s6 p""s2 p""{w p""{[7F] |""{r x""{2 p""q" p""q" P""q"
Same goes for sprocket alignment as is (wrong) and corrected 4/s/3 layout
well done checking though, that was outside my wheelhouse
Also not obvious how it might apply, but /code had all the text on grid paper. I guess it feels like it might be good to overlay something on our grid, but I don't really know what we have.
I did have a thought that the stray gray dots could overlay onto something to select characters or something like that
What were the dimensions of the code image?
entire image: 1600 by 1000px
split up image: ten by ten grid of 160x100px
individual binary squares: 20px by 20px
the graph paper looks to be about 30 columns and 39 rows but I could've counted wrong
Supposedly 1212ร742, but the grid paper itself isn't nearly that large.
Ight no the graph paper has 25 columns before turning into one more long row(s)
The bottom almost looks like long cells too.
I have a feeling that image is a hint and might not contain any actual information beyond that.
I think we're trying to go way too deep with a lot of things
Yeah, I have it, one sec
The Cyber Swiss Army Knife - a web app for encryption, encoding, compression and data analysis
The background bytes are the signpost
La simplicitรฉ est la sophistication suprรชme
One of my top 5 people of all time
one of the people of all time
We donโt exactly have a reference on puzzles yet to know how deep we should go tbf. But I 100% agree that we are probably looking too deep tho
I mean, they even made a post about how we were being too complicated
I donโt think itโs gonna be some crazy technical solution tho
"Strange loops" are definitely a theme here and we're in one ๐
judging by the way we over-complicated the poem, chances are high we're doing the same now lmao
Most definitely ๐
I'm still on 094316, but I have no idea what to do with it
I know itโs been posted before, but why is the url for the password page 094317, not 094316
I tried it
oh, misread, yeah
I'd say maybe there's a page with the right timestamp, but we know all the pages all the time ๐
as a timestamp, it goes to March 23rd, 2158 ๐
sorry, March 24th
Hmm? I thought it went to January 2 1970
? Is it not just the time of the security log
09:43:16 AM
the thing about that is, the actual memory bloc for that post is 5939823600 which is May 30th, 2026
And from what we've seen, usually the slugs contain the timestamp format for excel. So when using that same format, we get the year 2158
Oh, I see
and a TON of dates are like this, way off in the future for the sec-logs
They're usually unix epoch timestamps I thought
the funny thing is, the sleep logs contain the correct date in excel format
which establishes the baseline
once we get into sec-log world, all the excel dates are way in the future
where does that number come from?
I've been updating the posts spreadsheet with a new tab for tracking this insanity:
https://docs.google.com/spreadsheets/d/1mUOrEQqj5SqazYnwB9yjaRHICkJJVo0lJHHPSd9DbV4/edit?gid=1085556499#gid=1085556499
sorry that was a mistake actually lol
I was going off the date listed on the post itself, and in the full security log
1780134180 instead then?
correct
Ahh, it's the memory bloc numbers that are unix timestamps
yes
and the slug numbers appear to be excel timestamps
which is what all the sleep logs established
as for what to do with these discrepancies
no idea lmao
but it's there and we're tracking it
what is meant by excel timestamp?
I would recommend adding a column for unix conversions just in case a pattern comes out
I have two unix conversion columns
unless I misunderstand
The numbers after the security log is the time it was posted with a small error generally, but it matches the HH:MM:SS of the post?
unless im misunderstanding
I assume it's conflated from the excel standard dd/mm/yyyy hh:mm:ss
A column for "what would this be if it were a unix timestamp, not an excel timestamp"
I haven't understood the purpose either
HH:MM:SS appear to be omitted since all the times evaluate to 00:00:00
yeah I have that, it's the first Unix Epoch column
Where is Jan 2 1970 in this?
didn't include, since there's no excel timestamp in the URL
That's what I mean. There's a number that you convert into all kinds of things, but you don't convert to unix
I guess I'm not following what you mean haha
You convert what the excel version of that timestamp would be in unix, but not what it would be without the excel step
This
No worries at all
Unix is seconds past Jan 1, 1970 btw
Yes
So low number will always be in 1970
But just in case there is some sort of pattern. The first post is jan 1, 1970 in text, so maybe that matters
But it could just be a unix system resetting
because it has timestamp 0?
If we are thinking the interpretation of these damn pixels is simple, I can't stop thinking of how much this bottom row looks like an A
Jan 1, 2 and 3, 1970 lol
I think the whole image looks like galaga
Steering away from this for a second, we know our current cipher doesn't support diacritics (without extra hops)
I saw you say that, but how?
Just the multicolor kind of pyramid shape
I guess they are too high up in Galaga
I don't think that's the intent, FYI, just reminds me of Galaga
If we are placing stock in the words of the clip, perhaps we have a subtle indication to not be concerned about input using diacritics (just a side thought)
Maybe more Maybe Space Invaders?
Good thought, as the cipher ignores them
Like I could just be losing my mind at this point, but what if its as simple as just what letter do these pixels look like (not that it'd be easy/clean to interpret) but would be simple
I think anything that is an actual intended answer will be clear when we see it, not just a guess
Objective, not subjective
i had to stop looking at that image
it started looking back at me
I'm definitely reaching that point
I think I see what you mean ๐ฌ
That's it! When you look too deeply into the Void...Void Mother...GLASSSSSSS
It was a restless sleep last night
Those are some weirdly shaped boobies
what ARG's will do to you:
@turbid eagle Thanks for the column. Now that I see it, it's probably pointless ๐
blue footed boobies
you never know...
wouldn't it be wild if all this occurred over the span of a few days in the 1970s, and was somehow broadcasted 56 years into the future?
Emails between french people in the 1970s would be a bit wild ๐
Wild shout, do we know of any passwords/passkeys in-game with relation to the atlas?
16? Can't think of anything that has a password
Im more space invaders type
there are those little puzzles at the outposts and what-not, but that's about as password-y as things get I think
oh yeah, shifting left or right
those are always just a 123 - 231 -??? tho
(the answer is 312)
good job! ๐
I tried a few things related to em like Atlas, Hive, Glass etc, as well as Emily, loop16 and even the phoenix hex on the atlas passes, nothing worked
I discover a set of coordinates within the stars
literally 15 minutes over that hill
Regarding this, the timestamp links to TR4CE's original post. This one's got the missing 2 anomaly. Did anyone check if the full sequence of those missing bits works as a password?
Oh the pm is absolutely banging their head watching us all break out DQT tables and whatnot if the solution is intended to be simple ๐
They're probably like "what the fuck? I didn't even know you could do that!"
"breach attempts" caught my eye, but none of the language here's looking promising
They know because they did it, but yeah
Unknown had the 2 byte overflow from manual editing
BUT that doesn't mean it's worth jack
Because it all isn't lmao
Just means they're messing about with them
It's why I am starting to be sceptical of this image even being a part of the puzzle versus the video
The problem is that have we gotten anything from the image? They didn't post it just for aesthetic purposes
They did with the others though (location aside)
Yea it'd super weird to have a straight up teaser for this image half a month back and then have the actual image be useless
Which? There are a couple I think still have clues we didn't get
I think the one thing this image does tell us, is that it's not the memory we need to add when we do get the opportunity (due to looping ERROR)
Why is it posted on another platform instead of Wordpress itself? unless Wordpress doesnโt allow png hosting and that was necessary for this image
this might be an even more corrupted version of the teaser version
Is it the same image? I thought the RGB didn't match up
sort of like the way the poem became more and more corrupted until we finally figured out what the PM wanted lmao
Oh that's a fair point. They've had this account set up since may 9th, so they planned this out, but we've had images hosted directly on the site before too iirc.
the aspect ratio matches perfectly though, scales 1:1
Yeah it is def the Reddit image
I'm pretty sure TR4CE originally only had the image, before it was turned into a full page
maybe they didn't add it to wordpress so we couldn't track any changes it undergoes
That too
Can the uploaded image be changed? I thought they'd have to delete and reupload on the new site.
It's definitely weird especially since it's not like we had to solve a cipher or look deeper to find the link to the image
What if it's just as simple as assigning it the correct placement in the memory blocs to clear up the image?
also does anyone know if the string (C3q5NYF) at the end of the URL is auto-generated by freeimage.host? or if the uploader had to do that to their file?
fragment-1778433240.C3q5NYF
looking now in photoshop. Yeah, these are meant to be either the same image or very similar
The original had more data though
Itโs the original image name according to the metadata (unless free image host overrides that)
There's a lot of red missing in the newest image
Ah, so this isn't the real version yet either
And the red is slightly offset
But it is the same? They look so different to me lol
and green
C3q5NYF.png i mean
It's the same yes
I'm just doing ctrl+tab back and forth and comparing ๐
that's why I am led to believe this newest image is an even more corrupted version of the initial one
Wild y'all .. absolutely wild. Lol
A lot of the green is now cyan
But the original didn't have this clear ERROR background
So how would you propose cleaning up the original?
What if the extra detail was added by the corruption and it's supposed to be lower res like the new one?
(not being snarky genuinely curious as I don't know squat about fixing images and what not)
Here's the two layered, if it helps anyone
just tried it and yea, it's random
I think we've already dug into this no?
They don't match up, but they are very close
Cycle continues
Looks like a smear to the right
The CDAD/shadowcheck/whatever doesn't quant down to the same colour space due to edge artifacts and other things
Meh, I'm trusting vector to notice anything in images before I do ๐
If you exclude edges via thresholding the only seemingly signifacnt thing is the red
But a reconstruction path seems unlikely
It's probably better to think of them as sibling images
Yeah, they really look like two different images, just extremely similar
They're "the same"
The image we have is the original, the one that was posted to reddit was after it's been tossed in the microwave for a bit
Then the freezer, then the air fryer
They're a game of phone kind of idea
But you can see the binary background in the reddit image
then freeze dryer :p
They also harken to the hat problem/panopticon theme (loosely)
I assume the new image was attempted in audio? The original we got some things that were kind of like numbers but not quite. Maybe more clear in this version?
Got a whole bunch of samples in the repo
Hat Problem? Is that where I wake up with sleep paralysis and see a shadow problem in the corner wearing a hat and trenchcoat?
Still stand firm on that being auditory pareidolia
Of course he does ๐
When my paranormal/horror hobby crosses over with an ARG
Its possible that the password is separate from the image, and whatever is in there is what helps us decode the info from the image
where else could the password come from I wonder..
Yeah, I think the image is a clue, not the source of the password
A lot of green, at least
ive tried every interesting looking word in the trace sec log and full log, no dice so far
I'm guessing TR4CE is moreso tied to the image than the password rn
I don't think it would just be "use this word that is listed, verbatim". I think we have to figure out the right thing and then transform it, possibly with the cipher
watch the final password require rot 16 for no reason :trollface:
Or fibonacci sequence typed in reverse order
It's just the last 10 digits of pi, yall
completely* unrelated, any of y'all read qntm's stuff before?
this made me think of this: https://qntm.org/responsibilit
man
sometimes I type a password in
and my browser takes an extra couple seconds to load
and my heart skips a beat
like CONGRATULATIONS on another attempt
something that cracks me up about this ARG story line
is how we got access granted
only to be hit with a password page we can't get past ๐
I'm in.
Now I just need a password.
I was complaining previously that having DMs be the interaction meant it was not professional. They solved that problem!
They got fed up with us filling up their feeds and DM with attempts and dumped a password page on us lol
Time to do my waver part. Then i will go back to head wall smashing
one hour fyi
Ez
It would be funny if the password page didn't actually do anything other than log attempts and the PM is just going through the logs manually to see if anyone gets it right
or just logging to see what we're trying
Hah, they don't have an answer set, they are just watching our responses and picking what seems like the best answer ๐
"how did they come up with BobIsTheGOAT?"
The requests forward to wp-login, so there's definitely a solution set already
