Elektrine lite

← Feed

@tomw@mastodon.social

Post #968523

2023-02-04 22:25 UTC

People worry a lot about losing knowledge — about "burned-down libraries". Comparatively few people seem to worry about what happens if you take a billion books full of auto-generated, often-untrue junk text and *add* them all to the library. In theory, nothing is lost. In reality, everything is lost, because nothing useful can now be found.

Replies (40)

  • @tomw@mastodon.social there have been more than one philosopher of knowledge that has instructed to consider the backlog of things necessary to catalog in a theoretical library as a component of knowledge. Or rather as an impediment to knowledge.

    Open ##3219784

  • @ipd@universeodon.com 2023-02-05 00:41

    @tomw@mastodon.social That sounds like googles these days

    Open ##3219785

  • @acb@mastodon.social 2023-02-05 01:13

    @tomw@mastodon.social The Garbage Heap of Babel

    Open ##3219786

  • @doug@towns.gay 2023-02-05 01:23

    @tomw@mastodon.social every library is the library of babel

    Open ##3219787

  • @Phracker2Art@mstdn.social 2023-02-05 01:36

    @tomw@mastodon.social It actually gets worse, because the auto-generated junk text will be more optimized for search engines than anything written by humans, so pretty soon it will be impossible to find anything on the Internet outside of auto-generated junk and sponsored or corporate content. This genuinely worries me as a creator of blog content who understands the importance of SEO.

    Open ##3219788

  • @Cher@eldritch.cafe 2023-02-05 01:41

    @tomw@mastodon.social Ah, the Fact Sphere Gambit.

    Open ##3219789

  • @Whirl_Up_Sea@witches.live 2023-02-05 02:16

    @tomw@mastodon.social @BlackAzizAnansi@mas.to Borges’ Babel was a warning, not a suggestion!

    Open ##3219790

  • @JoshuaACNewman@hackers.town 2023-02-05 02:19

    @tomw@mastodon.social @thedarktangent@defcon.social On the other hand, this is a reverse lookup problem. A phone book with fake numbers in it doesn’t keep you from finding the right numbers. You just can’t be guaranteed that dialing any number will get you what you’re looking for. The concern with ChatGPT as an author is only when it fabricated false historical documents to cite — a problem already addressed in normal historical studies.

    Open ##3219791

  • @inthehands@hachyderm.io 2023-02-05 02:19

    @tomw@mastodon.social @BlackAzizAnansi@mas.to Zeynep Tufekci spoke passionately on this topic regarding Russian disinformation. Her argument was roughly that we should regard “flooding the zone” as a counterpart to censorship that achieves many of the same ends.

    Open ##3219792

  • @yunchtime@wandering.shop 2023-02-05 02:22

    @tomw@mastodon.social the signal to noise ratio is clearly something that Q-anon and its cohort of delusional cult ravings is designed to exploit. unfortunately for us the 'don't be evil' index that can winnow some of this chaff out, doesn't live up to its credo. so here we are.

    Open ##3219794

  • @donmelton@mstdn.social 2023-02-05 02:30

    @tomw@mastodon.social This. 💯

    Open ##3219795

  • @ML2@akkoma.sandhill.social 2023-02-05 02:32

    @tomw@mastodon.social I've been using a lot more site specific searches and even specialized search engines like marginalia.nu for exactly this reason. It also helps that my main search engine isn't Google.

    Open ##3219796

  • @tomw@mastodon.social @chancerydaily@masto.ai I have a friend who works at the Internet Archive and I told him recently that their collection has become instantly more valuable because the imports are all time-stamped; they will be the only large digital source of information uncorrupted by plausible AI.

    Open ##3219797

  • @tomw@mastodon.social you mean the internet?

    Open ##3219806

  • @davidslifka@mastodon.social 2023-02-05 02:44

    @tomw@mastodon.social Steve Bannon would call that “flood the zone with sh*t”

    Open ##3219807

  • @PixelRefresh@masto.ai 2023-02-05 02:49

    @tomw@mastodon.social Garbage in, garbage out.

    Open ##3219808

  • @tomw@mastodon.social I remember seeing a reference, I believe in one of Charles C. Mann's books on pre-Colombian America, about how some late Mesoamerican ruins show that at some point they began using their glyphs in nonsense ways, as if for the art rather than the actual grammatical or syntactic meaning. This is far outside my area of expertise, and it is possible that any parallels are apocryphal rather than directly related.

    Open ##3219809

  • @mackrauss@infosec.exchange 2023-02-05 02:56

    @tomw@mastodon.social There is a reason we have and will continue to need great librarians

    Open ##3219810

  • @garo@social.lol 2023-02-05 03:06

    @tomw@mastodon.social I cannot wait. ugh. ☹️

    Open ##3219811

  • @robotdiver@tech.lgbt 2023-02-05 03:06

    @tomw@mastodon.social The era of knowledge death.

    Open ##3219812

  • @mekkaokereke@hachyderm.io 2023-02-05 03:08

    @tomw@mastodon.social I suspect that writing certifiably produced before 2021 will be considered differently than writing produced afterwards, at least for some time. Like low background steel for writing. 🤷🏿‍♂️ https://en.m.wikipedia.org/wiki/Low-background_steel

    Open ##3219813

  • @tomw@mastodon.social @ncweaver@thecooltable.wtf This is the foundation of modern censorship. Not of stopping the flow of information, but of making the information difficult to find, burying it in mundane or contested or conflicting information. Of flooding the search indices and resources.

    Open ##3219814

  • @mbrewer@kolektiva.social 2023-02-05 03:20

    @tomw@mastodon.social I think this was the premise of the third edition of the cyberpunk RPG, cyberpunk 203X.

    Open ##3219815

  • @grvsmth@lingo.lol 2023-02-05 03:35

    @tomw@mastodon.social Thank you! I've been trying to warn people about this, but I haven't been getting through. Maybe they'll understand your formulation!

    Open ##3219816

  • @jstevenyork@mastodon.social 2023-02-05 03:44

    @tomw@mastodon.social We're getting to that point with the internet as a whole. All the information, and misinformation, in the world, at your fingertips, and no way to be sure which is which.

    Open ##3219817

  • @tomw@mastodon.social Holy shit I never thought of that. It's fucking depressing. Can we make human libraries exclusively for human-generated content?

    Open ##3219818

  • @AGTMADCAT@infosec.exchange 2023-02-05 04:26

    @tomw@mastodon.social That's what librarians are for! Here's hoping we have enough of them to weather the storm.

    Open ##3219819

  • @AllFallDown@mastodon.social 2023-02-05 04:27

    @tomw@mastodon.social long live books!

    Open ##3219820

  • @vga256@dialup.cafe 2023-02-05 04:19

    @tomw@mastodon.social this is one of the core theses of jane jacobs' (underrated!) book 'Dark Age Ahead'. i've thought about it many times over the past 20 years, and it has only become increasingly true over time.

    Open ##3219821

  • @tomw@mastodon.social especially if this happens in tandem with librarian jobs being cut / people being driven out of the profession due to harassment and low pay.

    Open ##3219822

  • @Vixen@bark.lgbt 2023-02-05 05:13

    @tomw@mastodon.social and this is why the internet is so flawed

    Open ##3219823

  • @s@stuartpb.com 2023-02-05 05:45

    I've been thinking about this a lot, calling it "the Infopocalypse"

    Open ##3219824

  • @ssfckdt@mastodon.cloud 2023-02-05 05:51

    @tomw@mastodon.social "explain the history of Google in two sentences"

    Open ##3219826

  • @Buggerlugs@mstdn.party 2023-02-05 05:51

    @tomw@mastodon.social in engineering it's called "Signal to noise ratio" or SNR. To extract signal, it has to be discernable within the noise. To suppress signal, increase the noise. The voice of truth can be suppressed by the babble of lies. The conspiracy theories aren't dangerous in themselves but pile them deep enough and people lose touch with reality.

    Open ##3219827

  • @roufamatic@hachyderm.io 2023-02-05 05:58

    @tomw@mastodon.social from Too Much Coffee Man, years ago. The boogieman back then was the govt, but the idea is the same.

    Open ##3219831

  • @Kazii@bark.lgbt 2023-02-05 06:00

    @tomw@mastodon.social it reminds me of how there is a digital library that has every book that could ever be ritten. a combination of your home town and your full name in order but no one will ever find it cause its just a bunch of random. https://libraryofbabel.info/

    Open ##3219832

  • @codingquark@mathstodon.xyz 2023-02-05 06:02

    @tomw@mastodon.social @float13@hackers.town even in theory things are lost because (information) entropy provably increases.

    Open ##3219833

  • @goose@honk.cozysumo.space 2023-02-05 06:07

    @annaleen@wandering.shop @tomw@mastodon.social I feel like that's a Borges story. 🤔

    Open ##3219834

  • @kingmob@kolektiva.social 2023-02-05 06:22

    @tomw@mastodon.social We will build igloos to live in blizzards of bullshit.

    Open ##3219835

  • @dhobern@scicomm.xyz 2023-02-05 07:01

    @tomw@mastodon.social There is a similar risk for every field that increasingly relies on citizen science data. Biodiversity research relies heavily on field observations to document the occurrence of species. This can have implications for conservation and land use and hence be of interest to extractive industries, etc. It seems inevitable that we will start seeing fake AI-generated observations that seek to obfuscate actual distributions of threatened species and communities.

    Open ##3219836