First Sundar Pichai, now Tim Cook, both admitting that generative "AI," by dint of its very foundations (cf. "Bias Optimizers"), will not and cannot stop bullshitting (cf. "On Bullshit Engines").
futurism.com/the-byte/tim...
Look at me. Listen to me:
If someone is selling you a "search" & "knowledge" "AI" product they claim is smarter & better & faster than you, but it literally cannot be guaranteed to give you simple true facts about consensus reality when asked, then that product DOES NOT WORK and SHOULD NOT BE USED.
First Sundar Pichai, now Tim Cook, both admitting that generative "AI," by dint of its very foundations (cf. "Bias Optimizers"), will not and cannot stop bullshitting (cf. "On Bullshit Engines").
futurism.com/the-byte/tim...
Use generative "AI" to spitball, hypothesize, overcome the tyranny of the blank page? I mean if the environmental costs weren't astronomical & the training corpra weren't largely stolen, then yeah, sure.
Use it for facts? Knowledge? To fully *Replace* thought, feeling, & creativity? Absolutely not.
And yes, the use of algorithms & machine learning to arrange search results based on profit motive has been a big problem for consensus knowledge for a long time, & everything here ↓, Safiya Noble and others long ago first wrote about as applying to said parameters of search. And LLMs make it worse.
Look at me. Listen to me:
If someone is selling you a "search" & "knowledge" "AI" product they claim is smarter & better & faster than you, but it literally cannot be guaranteed to give you simple true facts about consensus reality when asked, then that product DOES NOT WORK and SHOULD NOT BE USED.
Since this has escaped containment, here's a note that "Cf." is an abbreviation that means "compare," also often used to mean "see also." So when I say 'Cf. "Bias Optimizers" … "On Bullshit Engines,"' i'm not just throwing out phrases; I'm directing you to compare what I said to two Specific things.
First Sundar Pichai, now Tim Cook, both admitting that generative "AI," by dint of its very foundations (cf. "Bias Optimizers"), will not and cannot stop bullshitting (cf. "On Bullshit Engines").
futurism.com/the-byte/tim...
We need an Internet Encyclopedia. Something which can be counted on to hew to fact, and which is vastly more complete than what passes for encylopedias now. If only there was some entity with a mission to serve the public good, and which had the resources to create such a thing... if only...
Amazon Kindle bookstore uses AI / algorithms/ stochastic parrot to summarize readers' book reviews.
Results are "Some readers thought it was cool. Some said book had fast action. Others said it was rushed and jerky."
Not exactly a recommendation for the book or the "AI" "reviewer".
Just get some friends or at least a collegial work/topic group if you want to hypothesize, spitball, get started on a blank page and find it hard to do on your own. Doesn't require destructive software.
The people creating this grift don't have those so they made this junk.
Plus feeding the brain by looking at other people's artwork, and really appreciating what comes from other people's minds is an AMAZING cure for art block.
But I guess more time looking at art means less time grifting and can't have that.
It drives me crazy that in the #genealogy where the #1 rule is to cite your sources that people are willing to embrace generative AI technology where the source of the so-called information is deliberately obscured.
GenAI also steals the spotlight from useful work like handwriting recognition.
I agree. Although ai transcription is a big help and timesaver.
I could do it by hand, but this is 10 times as fast.
Upload, transcribe, correct, ready. #transkribus
Also for full text search it is great.
Two things bother me about OCR/AI transcription.
1. We learn more from the document when we do it ourselves. It's easy to miss things if we let the computer do most of the work.
2. OCR transcription is horrible when no one does a clean-up pass (I'm looking at you, Ancestry). Quality control? What?
Transcribing it yourself gives you a deeper understanding of the document though. But sometimes I just don't have the time for it.... Only one life, in which only a number of hours can be spent on genealogy.
Without correction it is horrible. But full text search can still give you clues. I have uploaded about 50000 pages to Transkribus when it was still free. Had it automatically transcribed and still can find documents I wasn't aware of... But also realising that I am probably missing half of it all.
AI generated trees will look none too different than those by people who treat their ancestry trees like a points based game, clicking on the top suggested results over and over, whirlwinding their ancestors across Europe, or grabbing dead trees that can be shoehorned. It'll muddy the field forever.
Literally saw on ad on prime earlier from Dell, advertising AI to detect fallen trees on train tracks "and look at this futuristic emergency response team that ensures you get places on time" and I'm just...
First off "lol" at a functional train system in this country. But also not an issue for it.
Like, I'm preeeeeeetty sure Japan's bullet train system is doing just fine without generative AI nonsense that'd probably claim there's a woman tied up on the tracks instead of...Honestly this ad may have been AI generated, which just proves your point more.
it's primarily an existential threat in the way it's going to be used by terrible, stupid people to do terrible, stupid things
imagine what will happen when these people double-down on the lie of "artificial INTELLIGENCE" to the point that critical decisions come down to computerized Mad Libs
Sam Altman and his kind are waging a determined war against meaning itself
when he says "i am a stochastic parrot, and so are you," what he means is he wants us to think there's nothing going on inside our heads that isn't happening in an LLM
and LLMs have no concept of facts vs. fiction
i don't disagree honestly
a vibe these techbros give me is that they are literally incapable of examining their own thoughts; they just have something pop up in their heads and they assume it's unrestricted genius simply because it's their heads that it popped up in
By which I mean the bullshit machine shillers are purporting to arbit truth when they cannot. They are piloting out of control machines that are going to crash us all head first into a cliffside because number go up
The non-AI driven search result options have pretty much disappeared.
I’ve moved to DuckDuckGo, but frequently the answer I’m finding is not to the question I’m asking.
I'm still so disappointed with ddg. they weren't fantastic with their results before, but hopping on this bandwagon absolutely ruined them.
right now im just occasionally using ecosia, which. still sucks.
otherwise im just going directly to sites under the wikimedia foundation or trusted sources.🫠
I've never really bothered learning my way around Tor, since I would need to have a trusted source of links already, but. I'm feeling more & more like that's the only real way to get anywhere any more.
idk, doing the research to get from step A to step C, back to A, then to B…near impossible now.
Being a researcher and relearning how to do your job every few years I’d gotten used to.
Being a librarian and have to teach the public how to do research when search is poisoned is new territory.
Thank you. I had not realized that!
I understand why people are using Perplexity; it’s pulling from more reliable sources than Google does, but it still makes stuff up.
Finding answers from trustworthy sources is never going back to print books - they just don’t get published anymore - now what?