The Backrooms of the Internet Archive | Hacker News Reader

The Backrooms of the Internet Archive

blog.archive.org

722 pointspassing112 comments

Why are most of the images missing on that page though?

It's not the first time I notice this — sometimes some images or SWFs or other resources on a page would be just gone, with "hrm, the wayback machine hasn't archived this URL" with seemingly no rhyme or reason. It's just kinda odd that this article talks about finding an image and points you to a page where only 2 out of 14 images were preserved without addressing this.

I love this stuff. Some other greats:

Doot doot: https://youtu.be/ZYcHOEjGzPA?si=_TP90YX-p5PkKdim

Michaelsoft Binbows: https://youtu.be/yDzAAjzbV5g?si=A2HTyCM8IU2MJHVQ

> Michaelsoft Binbows

Freaky, I had never seen the image before. It got posted on a Discord server I was on yesterday, and than you go and post a YT video. Heh

I remember seeing this before. Curious how / who found the original in the Wayback archives? Didn’t see that mentioned in the article.

Here's a YouTube video covering the discovery process. Ultimately, one person found it, but they were part of a wider team piecing together many parts of a puzzle, including outdated phone image numbering schema. I found it to be a worthwhile summary that doesn't really assume the viewer has much previous knowledge.

https://youtu.be/-1EKIIM3ShI

The Cybershot is a point-and-shoot digital camera, not a phone. Though Android does still use the DCIM folder, that whole basic structure, for compatibility with things that are looking for photos on drives, SD cards, etc. I think Sony phones even still use the DSC filename prefix for their photos.

Maybe you want to re-confirm that. In the late 2000s Sony Ericsson had a line of phones that had a focus on (then) high end integrated cameras called the C-Series, the C standing for Cybershot.

The photo was taken in the early 2000s though!

It seems there is actual standard behind it: https://en.wikipedia.org/wiki/Design_rule_for_Camera_File_sy...

Like many standards it's just formalized documentation of existing practice.

How did people hunting for the origin of this image discover the random niche website preserved by the Internet Archive that this image happened to come from?

The original URL of the photo was actually found on Twitter, where it had been posted in 2011. Wayback Machine was used only for the final confirmation. It's curious that this is not mentioned in the article, but I suppose it ruins the narrative.

I read about the whole thing last week at 404media, via waxy blog, which is a much more comprehensive article: https://archive.is/sj846

That is actually kinda fascinating given that it's directly in opposition to this blog post. I wonder who's telling the truth?

The legend lives on :)

The people who found it are telling the truth. The trail of discovery was 4chan then Twitter then Wayback Machine.

This seems very unfair. There's absolutely no point, anywhere in that IA blog post, that says "We found it". Anywhere! They're just providing information on the history of the file, from their archives, and detail into why it's an amusing story.

They even say

Naturally, as news of the Backrooms being “found” travels throughout the world, responses have wildly ranged. For some, this is a proof that “with enough eyeballs, all problems are shallow”.

How is that taking credit?

I think it’s usually pretty customary to attribute the original source if you’re going to write an exposition. They go through enough work explaining exactly the location of the furniture store, you’d think they’d have the courtesy to link to the tweet which actually made the discovery.

As has been pointed out though, IA has a pretty tenuous relationship with the Musk/chan adjacent parts of the internet, so it doesn’t surprise me they deliberately left those facts out.

You say that as if anyone knows.

From other comments in this thread, it looks like some guy on X solved it in 2018, yet others had no idea, and solved it independently too.

How many others "solved it"? Where is the official record of this? Who really was first? Did someone solve it before 2018? Did more than one party solve it, independently, recently?

Trying to untangle that and be sure, isn't simple.

They covered this by saying many eyeballs solved it, and then went on to their key message, highlighting the importance of archives, and showing how their archive proves the solution is correct.

My point is, pointing an accusatory finger at them, and saying they are trying to take credit is not fair.

Up until last month, the earliest known post/repost of the Backrooms image was an archived 4chan post from 2018, but it was believed to have been taken in 2012 or earlier based on the filename. So people have been looking for earlier posts/reposts of the image for years in an effort to uncover its origin.

During the recent successful search, the searchers trawled 4chan archives for early-2010s posts with similar image metadata to the 2018 Backrooms image copy. These archives were missing the original image files and thumbnails, but still retained some image metadata that could be filtered on (dimensions, image file md5s etc.) One of the searchers came up with a list of posts which might have originally included the image file, based on image metadata and context. Another searcher plugged the image md5 of one of these candidate posts (an April 2011 post recently added to an archive) into other archives, and hit on a post with a thumbnail matching the original Backrooms image from March 2011. At this point they'd finally found an earlier copy of the image, after years of searching.

Soon after, one of the searchers plugged the filename of the March 2011 post into Twitter's search, and came up with a post from 2019 which included the physical address and a link to the image source (this Twitter user had already found the source before the search had really begun, but it had gone unremarked upon at the time). The website had been replaced with blogspam in the interim. A searcher plugged this domain into waybackmachine and found a page with the image and a full explanation (it was taken during the renovation of a commercial property in Wisconsin).

Post from one of the searchers here: https://www.reddit.com/r/backrooms/comments/1d3pkif/how_the_...

this poor guy keeps getting ignored. here is the tweet https://x.com/rkfg_me/status/1130028610700664832

Should have posted that info in a sane location where people discuss these things instead of tweeting it into the void hoping that someone was looking.

Wait so, the Internet Archive was not involved at all in finding the original, but since the image exists in the archive, IA have written a blog post claiming to be crucial to its discovery? Seems like taking credit for something they didn't do to be honest. They didn't even mention the Tweet in the blog post which was essential to finding the image, which makes me think they want that part overlooked.

I think it's part of the recent trend to not mention Twitter/X because of its owner.

Any evidence for this or just mentioning it because you can?

I try to not mention Xitter because I just want it to go away. It was terrible as a conversation format before, and it's completely unusable now. Oh yeah, the owner is a raging douchecanoe too. But mostly it's just broken.

That wasn’t the context of the question. We are well aware that there are randos out there who don’t like Elon. I was asking why the GP felt the need to share the assumption of the author’s reasons for not mentioning the source, despite having no evidence. You’ve managed to lower the bar even further in the conversation, however.

dang changed a twitter link to non a twitter one in the past month or so because paraphrasing the twitter one may not work, let’s not prefer twitter, blah blah [nb original link worked fine].

Sorry I didn’t archive this but should be in the history.

That said I know many here have an involuntary eye spasm episode with Elon being mentioned but not sure IA does so not sure I agree with the original accusation.

They most certainly did not take credit, where did you see that?! I think this accusation is unfair.

In fact they say For some, this is a proof that “with enough eyeballs, all problems are shallow”. That doesn't sound like "we solved it!".

Instead, they're using this to highlight how vital archives are. It's a valid point, and there's nothing I see wrong with it.

I didn’t read the details of how they did it, but it would be cool if the Internet Archive exposed some kind of image hash / perceptual hash / similarity metric database, so that this task could have been a quick lookup in such a database.

I have often thought that it would be nice if the Wayback Machine had a reverse image search feature.

There’s actually a wonderful little mini-doc on YouTube that just came out the other day, produced by one of the people involved in the sleuthing:

https://www.youtube.com/watch?v=-1EKIIM3ShI

Almost all good memes seem to be extracted from some random niche website.

It wasn't needed for this but it would be great to have a fulltext (and binary!) and reverse image search of the internet archive. Often you know something about what you are searchng fore but have no clue about the location.

Of course this is not going to happen with the current resources of the IA and if it did it would probably just result in them getting hammered with DMCA requests and other legal demands for content that the "owners" didn't even know was on the archive.

I was sort of expecting an article called "the backrooms of IA" to tell me some interesting things about the unseen parts of IA. But it's still cool to see where the backrooms image came from. Presumably the building is still around and the hobbytown store seems to be in business as well.

Some videos I found of the store from 2014 mention it's the source of the "backrooms", so maybe this wasn't such a huge mystery to some people. Funny how something mundane like a hobby store in your town could become a world wide meme phenomenon.

You can visit the Internet Archive in SF. It’s worth it. They use an old church

Not everyone is near SF so an online article would still be appreciated.

Gotta love the found footage videos' photorealism and camera effects too. One of the reasons why a "retro" motif is commonly seen in these videos is to make it more convincing.

What found footage video do you assume to be most convincing? And how do you think photorealistic found footage videos will be made in the future?

Adding those imperfections to the video in post allows your brain to fill in the blanks and make it look more realistic. If it were a 4K video the CGI would be a lot more noticeable.

I think the retro look will probably stay as I feel it's part of the aesthetic. But maybe in the future we'll have backrooms-style videos of the current times, and then I imagine the retro/vintage aesthetic will go away.

How do you think CGI for photorealistic found footage would be optimized? What new methods might be used?

everyone says the Kanepixels stuff is the best, but in my opinion, establishing a lore and a story with this big corporate overlord and science fiction stuff was a mistake and only erodes the mysterious and malevolent nature of the entire thing.

There's a youtube account called "mattstudios" who does his own take, and generally spearheaded the "poolrooms" variant of this genre, that is far more grounded and tries to present the concept from the point of view of an everyday guy who "noclips" into it purely by mistake.

for me, this is the best backrooms take, not just for how realistic the video artifacts are, but because of how it captures just a guy recording something on his handicam in 1997 and how he'd realistically act: https://www.youtube.com/watch?v=KenTOGFwLpU

Kane's spiritual sequel "The Oldest View" is kind of like this.

>This agnostic, wide-ranging crawl likely represented both the original source of the image

Why do they say it's likely that the person who first posted the image on the message board got the image from the Internet Archive?

I thought the same thing. If you look at the crawled page, it's only one of like 20 images that survived.

So either the crawl got lucky and saved the only relevant image, or there is survivor bias.

Then again, the actual crawl might be triggered precisely because the image was linked.

What do you mean there is survivorship bias? That the only image used is the one that survived? Or it survived because it was used?

Something I noticed was that all other jpgs in this site have a lager number in the filename, for example: www.hobbytownoshkosh.com/Dsc00348.jpg

So maybe the crawler that saved this webpage had a limit on how many suburls it would capture and it sorted by name and then stopped at around Dsc00161.jpg, which is the name of the image in question. Though there is a Dsc00164 that is lost so it seems kind of unlikely...

That might have been the 161st image taken by an old Sony camera.

Cyber-shot model names use a DSC prefix, which is an initialism for "Digital Still Camera".

https://en.wikipedia.org/wiki/Cyber-shot

The filename for sure comes from the camera. My point was that the crawler stopped there and did not pick up the other images on the page because all of them have a higher number in the name and it stopped at an artificial number of sub urls

My guess is because the image didn’t exist in other archives and it’s a very obscure site so why would someone have seen that? More probable they stumbled on something random like this in the wild or on Internet Archive?

Why would someone have been looking at an very obscure site on the Internet Archive? Why is that more likely than looking at it on the web?

Because the obscure site had not existed for years when the copypasta first appeared. So someone would have had to have found that obscure site and then saved the image for years before using it.

Honestly, both options (website or archive) sound pretty unlikely to me. I'm wondering if instead it was a third option: maybe the originator of the copypasta was the person who originally took the photo. It would make sense for them to remember the event and go pull a good image out of their photo folder.

Or a fourth. The world is filled with a myriad of people, and we all have weird hobbies and drives. I can see an 'artistic type' of person, seeing an image and thinking "Oh, that's grungy" or some other label to the room/etc.

And "collecting" it. EG, saving it.

Some people will stare at live ants for hours, others collect rocks. I can imagine a person out of billions, liking "weird, musty rooms" or some such.

But who knows, heh.

They have an index of all images. Maybe someone picked a random image (I don't see a Random button, but one could exist?) or happened to look at this index the day it was added, or just clicked through to 2002 looking for something nostalgic https://archive.org/details/image

The discussion is about Wayback Machine rather than the link you suggest.

I didn't know about the old meme, but the image made me immediately think of The Stanley Parable. Not surprising since TSP is probably a descendant of the meme.

Stanley Parable outdates Backrooms by like 8 years. But they’re all basically about liminal spaces.

TSP isn't really about liminal space, it's about narrative decision making and the consequences of trying to account and develop for that in video games. I suppose you could say that it uses the liminal space of a barren office to achieve an awkward atmosphere that is meant to make you question everything about it, but that's a really small aspect of the game as a whole.

Sure. But it can be about a lot of things. And almost the entirety of the game takes place in liminal spaces.

If we're going there, we may as well mention Portal which predates Stanley Parable by 6 years, and features liminal spaces with the same eerie feel - in the game, you wake up trapped in an abandoned laboratory that seems to go on forever, which on its own hits the liminal notes a bit, and then you suddenly find yourself breaking out into service corridors and navigating the backrooms of the facility. Portal 2 (two years before TSP) continues that theme, though I feel it overdid the scale a bit, to the point of breaking suspension of disbelief.

Portal and especially Portal 2 have a very different feel to them for me - too dilapidated to evoke the same feelings as the office maze in TSP. There is also that TSP goes for retro while Portal goes for retro-futuristic.

Most of my dreams take place in similar places. Often though they are falling apart — like leaking ceilings, etc. Always they are labyrinthian, almost always it is night time (although the frequent windowlessness of the places would not make that obvious). They are often populated though — maybe college-age students, sometimes more like a mall.

I've always been fascinated by the place of my dreams. I have asked around but haven't had anyone confirm having similar dreams.

When I was a teen (and a bit younger) I had night terrors. I did not know the name of them at the time (no internet yet). But sometimes they featured a room that extended so far in every direction that you could not see the walls. Something like an all-white parking garage, I suppose. I feel like a family friend a little older than me tried once to hypnotize me when I was young — they might have used a similar description: a room white that extends to far that you cannot see the walls. Perhaps that was the source of the imagery in the night terror dreams.

I wrote a game decades ago where you (well, a paper airplane) wander a seemingly endless house trying to escape. I am not sure which came first though — the dreams of an endless space or the game.

Wonky spaces seem to be a fairly common thing in dreams. One that I used to have every so often was walking around inside what at first seemed like a normal room, then looking up to see that the ceiling was several tens of stories high, as if the building were a skyscraper with only a single tall and narrow room inside. The unexpected height of the ceiling always caused an intense sensation of vertigo, causing me to fall backwards in my dream (and sometimes wake up IRL).

> I have asked around but haven't had anyone confirm having similar dreams.

Well.

I dream of several specific places that I have never been. One is a group of buildings featuring 3 or 4 bay garages in the desert, think New Mexico flatlands. The buildings are typical corrugated steel siding and roofing, painted in a light grey and surrounded by 10 foot barbed wire chain-link fences. It is always sunny and the heat is near unbearable. Sometimes there are men with large trucks outfitted in a post-apocalyptic manner. Sometimes there is no one at all. I always have the sense that the undead are nearby and that I need to be very, very careful.

This is one place that I come back to regularly. There are others that I don't recall as clearly.

I frequently have dreams where buildings from my past have rooms which have doors opening to corridors opening to rooms endlessly like a labyrinth. Also often populated, mostly some sort of school or shelter building. They spark fascination and a sense of infinite exploration rather than terror. I vaguely remember I knew the source of them (some sort of escape fantasy?) in my childhood.

I'm surprised that despite all the talk about liminal spaces, no one mentions the lack of windows in those spaces and how much they add to the creepyness.

I found this especially noticeable in Stanley Parable. Or rather, the offices there have windows, but they are all opaque, showing just a featureless white (and a few are mounted on interior walls, making you doubt whether they really are windows or just LED panels).

At least for me this had an enormous effect to the drearyness and general feeling of disorientation in the game.

I've been fascinated by the genre of creepy "liminal space" pictures for ages because the universal nature of how offputting they are seems to transcend a lot of cultural barriers

I think that it actually boils down to a very deep seeded similar animalistic fear all of us have: the agoraphobic open spaces with no place to hide, lack of furniture or shelving to hide behind either, unreliable lighting with stark shadows that could hide danger, no windows or inability to judge the passage of time or ostensibly escape through. They're always desolate with no people around, and what's more, evidence that nobody's been through in ages so you can't even hope to stumble upon anyone. Abandoned places, dangerous decrepit old buildings, places too large to count on passing by another person, places abandoned at night because people have left.

In the end, you feel trapped like a rat in a shoebox with nowhere to go. If something trying to get you appeared, you'd have no options to run or fight back, and nobody to call out to for help.

one of my favourite series of images were of an american suburb, rows of identical houses with fences, far apart from each other so interaction with neighbours was minimal, short little trees, so everything is exposed. Because the construction was so new, nobody had moved in yet, so there were no cars, no lit up windows, absolutely no life. In lieu of walls keeping you trapped, the sheer size was the isolating factor: you could run away for 10 minutes and get nowhere.

You might enjoy Vivarium if you haven't seen it already.

The notion of vast abandoned underground "backrooms", maybe with a couple people in hazmat suits on their way through also perfectly fits into the Westworld TV series. Though of course they didn't go with an 70's office vibe.

But the copy-pasta for the image builds on those influences, not the other way around.

This IA article is by the same Jason Scott of http://textfiles.com/jason/

See you all in the backrooms

Anyway, I'm sorry if some people are reading this hastily written blog entry to seem like the archive is taking credit for the process of discovery being done by people. The phrase likely does not mean definitely and perhaps I should have used a different word when I wrote the entry. But the fact remains that the wayback machine is the only place you can see the image in the context of its original website, and that is only happening because the archive is doing such a general crawl. That's all I wanted to get across, all hail the wayback machine, have a great day.

Thanks for writing the blogpost! I think it's perfectly valid as a fun demonstration of the utility of the wayback machine.

I found the article very interesting and didn’t for one second feel the author was ‘taking credit’ in any way. A few unreasonable critics are a good signal that you wrote something that was widely read and enjoyed. They’re not representative, they just comment a lot.

It wasn't needed this time but it would be great to have a fulltext (and binary!) and reverse image search of the internet archive. Often you know something about what you are searchng fore but have no clue about the location.

E.g. I'm interested of finding old installers of a particular type (identifyable by certain byte sequences) but they could have been hosted anywhere.

I guess this is not feasible with the current resources of the IA and if it did happen it would probably just result in them getting hammered with DMCA requests and other legal demands for content that the "owners" didn't even know was on the archive.

Immediately made me think of the set for the show Severance

If you think that this would be right up SCP’s alley, you’d be wrong, it’s not included (yet). Here’s a Reddit discussion on the topic: https://www.reddit.com/r/TheBackrooms/comments/bs2zog/why_th.... The SCP 682 referred to here is, of course, the undestroyable creature: https://scp-wiki.wikidot.com/scp-682.

Now someone should make a game where you design indoor rc car tracks in the backrooms.

Someone should convert an unused warehouse or shopping mall into a real-world backroom maze escape game.

There should be more things in the world like Meow Wolf - this is an aspect of that in person exploration experience.

I wonder if anyone has a list of that kind of space.

Relevant: <https://scp-wiki.wikidot.com/scp-3008>

Not sure about design, but for driving them you have re*volt

And, as expected, there is a backroom level: http://revoltzone.net/m/tracks/70429/Backrooms (There are probably others, this was the first result after a search)

These pictures make me think of silicon valley startups, especially after the dot-com bubble burst. Offices filled and popped, usually leaving trails of rack-mount and cube panel screws.

“ However the original, anonymous user stumbled onto this photograph, it appears it was taken from either the Wayback directly, or the Wayback Machine crawled the same site the user had found, and kept that webpage’s preservation for over 20 years.”

Something people used to do is to use search engines to find images with file names that indicate they were from a digital camera. You could find all sorts of interesting photos that people had uploaded to share with their friends not realising that search engines would index them.

I recall a thread years and years ago about a guy who found an abandoned digital camera in the woods, only to take it home and extract the photos from its SD card, and as everyone poured over them, they began noticing odd figures in the trees and so on.

That's the kind of fun group activity I love about these mysterious things

Got a link to those photos? Sounds like an interesting story, whether true or not.

https://creepypasta.fandom.com/wiki/I_Found_a_Digital_Camera...

The photos feel like a story being told. Meant to be in an almost serial way.

Yeah where's that link? That sounds fun.

You can actually do this with YouTube. By search for MOV001 etc.

Mental Outlaw did a video demonstrating this a while ago: https://www.youtube.com/watch?v=t_Ho_KDPi_U

Good luck “searching” for anything on YouTube in 2024..

It's 2024 and need a search engine that doesn't suck

https://kagi.com is by far the best out there IMO.

Also:

- http://startpage.com

(Quite the story, that ...)

http://astronaut.io/ - YouTube videos that have almost zero previous views

2019 https://news.ycombinator.com/item?id=20432772

2017 https://news.ycombinator.com/item?id=13413225

I expected dumb gameplays and influencers. Somehow zero of those. Almost all the random samples have been immensely interesting.

There is a subreddit for Open Directories!

There's an even more fitting (inactive) subreddit for filename searches here: https://old.reddit.com/r/IMGXXXX/

Here's a list of prefixes to search for:

    IMG XXXX
    MVI XXXX
    MOV XXXX
    100 XXXX
    DSC XXXX
    SAM XXXX
    HDV XXXX
    102APPLE 
    IMG XXXX
    FILEXXXX
    GOPRXXXX

There's also the excellent http://astronaut.io, which I found via Hacker News. It automatically plays short segments from YouTube videos with those default filenames, inevitably surfacing a kind of gallery of the mundane: The unproduced, untargeted, mostly unwatched, everyday moments of real life.

The amount of (AI generated?) crypto spam is mind boggling. Most of the other videos seem to be amateur sport events, but the rest contains a few gems of random stranger's life :)

I let it scroll through about 30 and didn't see either of these. I saw a couple small church services, some random snippets of people walking throughout various areas, what seemed to be an interview for someone's school project and some videos of pets.

This reminds me of opendirectories subreddit. So many random stuff.

Back in the day, searching for "top secret", "highly confidential", "for private use only", etc, ... would return interesting results.

Distinguished seekers also searched: "NOFORN" "EYES ONLY"

I love and appreciate the Wayback Machine, but using it is such a bittersweet experience. So many of the crawls are incomplete. I’ve managed to find pages that hosted content of interest to me, only to find that particular resource unavailable. And if it’s not on the Wayback Machine, it’s just gone forever. Feels like tracking an old friend down to their tombstone.

I quite like the fragility of it, it makes it more apparent that everything is transient. In a way I wish the IA had a half life on content, that it would decay over time, pages and images would be randomly deleted. Little by little it would rot and become nothing, a reflection of humanity.

I suppose that's the internet itself...

> everything is transient

I agree. Permanence should be tied to individual wills, not collective inertia. If you want a permanent thing, work for it, host it and publicize it

Sort of describing entropy yes? All things will decay unless external energy is continually applied to the system to maintain an ordered state.

Every future historian is hissing at you right now.

So much of the past is completely opaque to us because of decay, intentional destruction and lack of interest. I think there's a moral imperative to preserve.

Things that are important to society will naturally be preserved. I believe in the moral right to be forgotten.

We cannot, and should not, preserve all of knowledge forever.

Don't get me wrong, I love the internet archive, and the team behind it are incredible. As a resource it's very important to maintain and preserve. However, I'm sure that at some point, either due to the economics of it or through hardware failure, the content saved by the AI will begin to be lost. I don't see that as a bad thing.

If someone requests for their content to be deleted, IIRC the IA does it. In other cases however I don't see the need to remove old(er) content. Particularly also because older content/webpages were far lighter than modern equivalents - you may need to delete 10 or even 100 old website archives to store one new one.

Decay is natural, but so is the human/animal urge to stop it.

Yes. Both content but also the technology to display this content. I put this on my personal web page 21 years ago:

http://trondal.com/magisk/magic.html

At the time (or maybe a few years before), clicking this button would show a dropdown menu linking to a bunch of web pages. Now, the button doesn't work anymore and I think most of the links go to missing content.

> Feels like tracking an old friend down to their tombstone.

I did this yesterday. He went in 2016.

Definitely feels incomplete. It should at least make an attempt to capture videos from the crawl. It feels like it does less than what yt-dlp would do if given that URL.

Reply on news.ycombinator.com