Comments
I briefly touched on the topic of "doomsday archives" in a recent post. As far as I'm aware currently, the two most notable organizations working in this area are the Arctic World Archive, whose archive is located in Svalbard and whose headquarters is in mainland Norway, and the Arch (pronounced "ark") Mission Foundation based in the U.S. with archives in multiple locations, including on the Moon — really.
I'm personally interested in storage media that are long-lasting and resilient to major disasters and don't require a continuous electricity supply or continuous copying over to new discs or new hard drives. Paper is probably the best medium, all things considered. It's cheap, it lasts a long time, and anyone can read it without anything other than the paper itself, as long as they can read. The main disadvantage is storage density.
In the form of ebooks, you could keep 50 million books on fifty 20 TB hard drives in a closet for about $10,000. Storing 50 million paper books would cost many millions of dollars, both for the books and the building to keep them in. So, hard drives are unbelievably superior in terms of storage density. But they are incredibly short-lasting and non-resilient to major disasters.
New technologies are trying to combine the long-lastingness of paper with the storage density of hard drives. The Arctic World Archive uses piqlFilm, developed by its associated for-profit company, Piql, in Norway. piqlFilm involves printing QR codes on film, estimated to last 2,000 years under ideal conditions, along with normal text instructions for decoding the QR codes. (You can also just print text or black-and-white images on it, but that's much less information dense.) The Arch Mission Foundation uses Nanofiche, which is text and images engraved on nickel. It only requires magnification to read.
Microsoft is working on putting data into quartz glass. The research project is called Project Silica. The storage medium is a clear square piece of quartz glass about the size of a floppy disk or a glass coaster. Each one can hold 7 TB. The estimated longevity is over 10,000 years. However, it requires advanced technology to read the data. It isn't intended for a doomsday archive. Microsoft is exploring its use for cold storage for cloud data as a competitor to hard drive storage, on the thinking that rarely accessed data could be more cheaply stored on quartz glass than on idle hard drives.
Encoding data on dehydrated DNA is another idea that has received a lot of attention and is an active area of research. Dehydrated DNA can last a long time. DNA is also incredibly information dense. But it is ungodly expensive. It also requires advanced technology to read.
We are somewhere in between having no good options for storage and having the perfect, magical solution. We're somewhere in the middle. Paper is really excellent and might be the best option overall. piqlFilm is a very intriguing option, but it costs about $30 per GB and requires cameras and computers to interpret. I don't know the cost of Nanofiche, but I imagine it's expensive.
In a way, books and libraries are already fairly good doomsday archives. Particularly since they are everywhere. The main three ways paper could improved upon are:
- Storing data types other than text (with a rare black-and-white photo or diagram)
- Storing a lot more information in a small physical volume for a manageable cost, e.g. storing the equivalent of 50 million ebooks in a closet for $10,000
- Extending longevity even further than the few hundred years estimated for modern acid-free paper with a small alkaline buffer (although longevity depends a lot on storage conditions, particularly humidity, temperature, and exposure to direct sunlight — cool, dry, and dark is ideal)
On the first point, I think piqlFilm might be the current best option for storing recorded music for hundreds of years (or more). Vinyl records might also be quite good, although it's hard to find any reliable, hard data on their estimated longevity. Every source seems to just say about 100 years without any explanation of how that estimate was arrived at.
Movies and other video can be printed on film. Movie studios routinely store a copy of heavily digital movies like Avengers: Endgame on analog film as part of their standard backup strategy. Modern film, which uses the same polyester base as piqlFilm, apparently has good longevity and is expected to last for centuries when stored properly.
The part of doomsday archiving that is possibly more vexing than just storing the data is figuring out how to present information to a society in ruins that has suffered potentially decades or centuries of disruption, decline, or collapse from a major disaster like a nuclear war (God forbid). This is particularly salient when the storage medium requires building a certain level of technology to read it, which is the case for every medium except paper. But even apart from that, there are questions of how to present the information in a way that is understandable, that shows people where to start, and tells them what's most important. You can't necessarily rely on people in a future scenario like that knowing what to look for and if the pile of information is extremely large — e.g. if it contains millions of books — then it could be daunting.
If anyone wants to ask me random questions about this or related topics to see if my research has already turned up an answer for you, please ask away (even if you're seeing this comment a long time from now, although you may need to send me a private message that includes your email address).
This might be a dumb question, but shouldn't we be preserving more elementary resources to rebuild a flourishing society? Current EA is kind of only meaningful in a society with sufficient abundant resources to go into nonprofit work. It feels like there are bigger priorities in the case of sub-x-risk.
I’ve definitely thought about this and short answer: depends on who “we” is.
A sort of made up particular case I was imagining is “New Zealand is fine, everywhere else totally destroyed” because I think it targets the general class of situation most in need of action (I can justify this on its own terms but I’ll leave it for now)
In that world, there’s a lot of information that doesn't get lost: everything stored in the laptops and servers/datacenters of New Zealand (although one big caveat and the reason I abandoned the website is that I lost confidence that info physically encoded in eg a cloud server in NZ would be de facto accessible without a lot of the internet’s infrastructure physically located elsewhere), everything in all its university libraries, etc.
That is a gigantic amount of info, and seems to pretty clearly satisfy the “general info to rebuild society” thing. FWIW I think this holds if only a medium size city were to remain intact, not certain if it’s say a single town in Northern Canada, probably not a tiny fishing village, but in the latter case it’s hard to know what a tractable intervention would be.
But what does get lost? Anything niche enough not to be downloaded on a random NZers computer or in a physical book in a library. Not everything I put in the archive, to be sure, but probably most of it.
Also, 21GB of the type of info I think you’re getting at is in the “non EA info for the post apocalypse folder” because why not! :)
That was my first thought, but I expect many other individuals/institutions have already made large efforts to preserve such info, whereas this is probably the only effort to preserve core EA ideas (at least in one place)? And it looks like the third folder - "Non-EA stuff for the post-apocalypse" - contains at least some of the elementary resources you have in mind here.
But yeah, I'm much more keen to preserve arguments for radical empathy, scout mindset, moral uncertainty etc. than, say, a write-up of the research behind HLI's charity recommendations. Maybe it would also be good to have an even small folder within "Main content (3GB)" with just the core ideas; the "EA Handbook" (39MB) sub-folder could perhaps serve such a purpose in the meantime.
Anyway, cool project! I've downloaded :)
Yeah i guess that makes sense. But uh.... have other institutions actually made large efforts to preserve such info? Which institutions? Which info?
Huh, maybe not.
Might be worth buying a physical copy of The Knowledge too (I just have).
And if anyone's looking for a big project...
Another easy thing you can do, which I did several years ago, is download Kiwix onto your phone, which allows you to save offline versions of references such as Wikipedia, WikiHow, and way, way more. Then also buy a solar-powered or hand-crank USB charger (often built into disaster radios such as this one which I purchased).
For extra credit, store this data on an old phone you no longer use, and keep that and the disaster radio in a Faraday bag.
All done :-) (already had a solar/crank charger+radio). Thank you!