@tomw
That sounds like googles these days

@tomw The Garbage Heap of Babel

Incorrect documentation is often worse than no documentation.

Bertrand Meyer

https://www.azquotes.com/quote/709604

@tomw 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.

@tomw Ah, the Fact Sphere Gambit.

@tomw I miss webrings

@tomw @BlackAzizAnansi
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.

@tomw

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.

@tomw This. 💯

@tomw @chancerydaily 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.

@tomw I think you are against AI written stories in libraries. If this is the case I agree fully. I think it is already past time that AI generated materials should be permanently branded as AI created.

@tomw
you mean the internet?

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

@tomw Garbage in, garbage out.

@tomw 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.

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

@tomw I cannot wait. ugh. ☹️

@tomw The era of knowledge death.

@tomw

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

@tomw @ncweaver 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.

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

@tomw 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.

@tomw This is already happening in wikipedia, Pew Research CEO on a blog made a age rage to help their research into generation z but the person who made edits to the page never told anyone why such numbers were created. Many news outlets and government agencies used the fake age range of 1997 - 2010 has Pew research stance on Gen z. Now in 2023 there is no agreement on when someone in Gen z was born or when Gen z ended. This is causing many misleading and mistakes in data research.

@tomw Holy shit I never thought of that.

It's fucking depressing.

Can we make human libraries exclusively for human-generated content?

@tomw That's what librarians are for!

Here's hoping we have enough of them to weather the storm.

@tomw long live books!

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

@tomw and this is why the internet is so flawed

@tomw "explain the history of Google in two sentences"

@tomw 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.

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

@tomw 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/

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

@tomw 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.

@tomw There was a brief moment in time that you could trust the internet, but that was quickly eclipsed by the swindlers, hoaxers, and rotten.com :)

We've not been able to trust text on the net for a whiles, nor (photoshopped) images, now deep faked videos and AI generated data

It's been a race to find a way to share information in such a way that it was impervious to those who lie.

And we've been losing every time

@tomw
A Colombian judge used chatGPT in a ruling.

So chatGPT makes up answers out of thin air, those answers become part of public record, search engines index them, then you're done: no longer able to search for factual answers to questions.

https://www.theguardian.com/technology/2023/feb/03/colombia-judge-chatgpt-ruling

@tomw That's the whole point of disinformation, not to convince you it's true, but to swamp the truth in so many lies you can no longer find the truth or recognise it when you find it.

@tomw indeed, the principle of the purloined letter, applied at scale...

@tomw
This is one of those bad ideas I'm reluctant to share because you just know it will inspire the assholes out there.

@tomw

My friend Borges found this out the hard way. He's still searching for the accurate catalogue of all the books in his library 🫤🤷‍♂️

@tomw Here’s a question: are our libraries *already* filled with inaccurate information? How do we know? Who judges?

@tomw
In theory, this will happen *in the future*.
In practice...

@tomw this is one of my main concerns when it comes to AI text generation. Just imagine the disinformation possibilities.

@tomw @lpsrocks

(I appreciate people replying about actual books and libraries but this is, like, a metaphor y'know.)

@tomw
The vast majority of volunteer authors and proofreaders in Wikipedia try their best to prevent the addition of false, stupid or manipulative information.
But there are too few of us.

@tomw isn't this exactly what French intelligence did to render the political document dump unusable??

@tomw reminds me of Borges’ Library of Babel.

@tomw has this happened?

@tomw
... and false balance is part of the game 🤷‍♂️

@tomw Bannon's “Flood the zone with shit” at scale.

Maybe somebody can design an app that finds the useful stuff?

@tomw You are also describing advertising, except it isn't junk but worse: corporate misinformation and propaganda whose aim is to maximise profit at the expense of humans.

@tomw but if you added infinite auto-generated books we would get Shakespeare back. So there’s that. Just keep adding.

@tomw Keeping me up at night as we speak..

@tomw sounds like kindle unlimited to me

@tomw This made me realize a thing I wasn't able to understand before: people say that Stanisław Lem predicted the Web via a science fiction story about a Daemon of the Second Kind (producing an equivalent of Borges' Library of Babel).

I knew Web as of when spam was obviously recognizable and linking allowed finding a lot of good information. Now I see that 1965 story as being more about using ChatGPT-like tools to get a lot more useless information.

@tomw Every day now, I am seeing my somewhat distinctive name incorporated into random nonsense web pages. The frequency of these spam posts is now about the same as genuine mentions. If the plausibility of these junk pages continues to improve, it may soon be difficult to find genuine mentions of me on the web.

@tomw People out here trying to make the Library of Babel a reality

#Borges #borgesposting #library

https://libraryofbabel.info/

@tomw

I'm old enough to remember when we thought the internet was going to usher in a golden age of knowledge. So naive...

@tomw yeah but early people wrote books

@tomw This is the central theme of The Library of Babel by Jorge Luis Borges.

@tomw That is a SEO problem. A decent AI system with librarian support will fix that problem

@tomw it's already been done.

@tomw I get that it's a metaphor, and yes it points to a big problem internet-wise.

But libraries are curated. And the Florida outrage is about who should get to do the curation - teachers or politicians.

@tomw this is stupid but like, way back in college when our physics department wanted to throw out all the old physics books, I took every single one. I still have them in plastic tubs.

I had this absurd idea to bury them, like the Rosetta stone.

@tomw I was never convinced by the whole PGP signing web-of-trust thing that was (more) popular about 15-20 years ago. I wonder if we should bring it back, and use it for actual trust -- I'd be happy to mark pages I trust and am an expert on, and people could trust friends and particular experts, then (weakly) trust friends of friends, etc.

Not perfect, but we probably need something against the storm of AI generated nonsense.

@tomw +1 It's all about "signal to noise ratio" and dirty data.

The sticky point is the question of who gets to decide what is "untrue" and "junk"....

@tomw Another way of looking at this is that propagandists try to make *everything* a library so that you can 'do your own research' by watching YouTube videos from people who have no expertise in what they are discussing and would never be able to get an old-fashioned book published by any respectable publisher.

@tomw you sure all the books were true before?

@tomw That's called "the Internet", IIRC.

@tomw This worries biomedical scientists on a daily basis, which is why they've made incentives to have perverse ranking systems devised and sold back to us to help us rank one another in terms of "productivity", all of which systems, too, have been gamed or diluted themselves. #Pubmed #Clarivate #googleScholar

@tomw I have been noticing this about the internet for a few years now. It is getting very difficult to find useful information from among the AI-generated listicles these days.

@tomw
Beyond Borges and into the Dark Age Ahead ...
The majority of we poor primates are blindly heading towards mass extinction when overwhelmed by heat and associated climate catastrophes having been 'controlled' by the dishonest political, economic and media elites whose strategy of drawing public attention away from important issues by flooding the media with continuous distractions, deflections, 1/3

@tomw
I hate fear, but I fear people who hate knowledge.

I'm petrified of book burnings.
I'm anxious of library shutdowns.
I'm perturbed by those who want to censor, ban, and leave the people without eyes, ears, or a cranium.

@tomw I've seen internal document systems rendered completely useless by storing absolutely everything, regardless of how useful it is. Trying to search for anything just reveals garbage. I'm sure the real document I was looking for was there, because the person who created it saved everything, but it's just impossible to find.

@tomw Someone needs to develop a search engine to sort through the endless text being created by AI tools. You could call it the "Better Optimized Robot-Generated Exclusionary Search", or B.O.R.G.E.S.

@tomw if you want a preview of this, try searching for *anything* to do with cooking, baking, or gardening on DuckDuckGo. Entirely links to autogenerated, conglomerated crawled content, turduckened together with tons of ads

And because it’s readable and mostly plausible info, it often takes me halfway through the page to realize it’s junk, remember DuckDuckGo is unusable for these topics, and switch to google

@tomw

And now there's #ChatGPT

@tomw you just described Amazon shopping lately. 5000 alleged matches, no useful filter to narrow the search. (Try finding a 3/4” wide belt for example. China seems to think 1.25 to 1.5” is what all US men want.)

@tomw I see someone has read his Borges

@tomw You are right, even in the most literal interpretation of "library." Librarians license journal and ebook packages for our collections but can exercise very little control over their contents. I don't know how to fix it - #CollectionDevelopment is not my specialty - but I see this dynamic in the catalogs of the academic library where I work and the public library where I'm a patron. Lots of content that no one would buy on its own merits, drowning out high-quality material.

@tomw this is why libraries the institutions are so important, and different in quality from libraries the buildings. It's the job of a librarian to fill the library with information, refresh it, and remove (or archive, if important) what's stale.

@tomw
Hence my short story, "Noon in the Antilibrary."

https://www.technologyreview.com/2018/08/18/104097/noon-in-the-antilibrary/

@johnwinkelman

@tomw unless there were some way of making a list of the ones before the polluting. but in the absence of clear marking and/or heuristics to detect churned content going forward it is indeed a tragedy of the commons...

@tomw That's some properly good red team thinking!

@tomw
About 15-20 years ago, we stood on the cusp of the information age, but collectively decided that drowning in the trash of the data age was the way to go...

@tomw yeah we've been quite concerned about this, personally.

@tomw Well this is why we have librarians.

@tomw

This process started quite a long time ago but it's been accelerating steadily, much of it driven by the way Google ranks search results and pays sites for ads.

ChatGPT is just throwing cans full of gasoline on the existing fire.

@tomw
Hear, hear!

@tomw great point, so bad.

@tomw thanks for the new anxiety. Now I won’t be able to sleep tonight.

@tomw Vernor Vinge's Fire Upon the Deep novel features a malevolent AI bent on galactic domination; the messages warning of its spread are buried in a sea of conflicting and incoherent babble.

@tomw there is a really prescient short story along these lines, by Borges about eighty years ago. (“The Library of Babel”.) I think about it a lot lately.

@tomw

This is why open source software documentation has a reputation for being poor quality, IMHO.

There's actually good documentation for most of it, but also documentation that is obsolete, unofficial, etc. And that dilutes the quality that the searcher experiences. If you get the wrong source, it'll be subtly wrong a lot of the time.

One reason that it's really important to include version numbers and dates.

@tomw This is a subplot in the Neal Stephenson novel "Fall; or, Dodge In Hell". There (spoiler alert) for all practical purposes it destroyed the Internet.