the AI (ChatGPT) generated the error (and several others if you like carefully) and I thought it was funny so let it stand. the prompt [make a movie poster for “The Alienator vs The Deceiver”; the first guy should like Dario Amodei, the second like Sam Altman] didn’t ask for it.
The destruction of rare books is a Crime Against Humanity, the apotheosis of Greed, the Hydrogen Bomb to literary culture... Dario and Sam, as well as many other tech titans, are in the thrall of their toys, driven by the lust for power... "to have dominion." What if paintings and sculptures are next on the digitize agenda? They already are being plagiarized.
If folks like Amodei and Altman thought they could make a dollar off of the Mona Lisa and had to destroy it to avoid paying a single penny to do it in a nondestructive manner, they wouldn’t even hesitate for a nanosecond to do it.
The citation is about bulk buying from a company called isbndb. It’s worth noting that isbn numbers only go back to 1970 — so the books being cut up extend from 1970 to 2022. They aren’t antiquarian books ( which is what “rare” originally evoked for me ). No Mona Lisa’s or Aristotle’s lost treatise on comedy. “Rare” is a load bearing term in this discourse, as the kids say, and it is t clear it means what we think it means.
We fetishize the physical tactile object of a “rare book,” but the value is in the content.
Would you rather the rare book is non-destructively scanned and free to view for all the literate people of the world? Me too. But if I had a choice between destructively scanning and non-destructively scanning I guess I would prefer the book be destroyed if its actual contents could be liberated. Has anyone found out if Anthropic reasonably had or has that choice? Non-destructive scanning as practiced in the past by Google took decades, and I cannot really imagine what kind of human page-turning labor they exploited for that worldwide ingestion. Meanwhile, it is safe to say that most of the content scanned by Google has been hampered by understandable legal complications and is effectively locked away.
Will Anthropic now provide the complete scans in their entirety, free of charge, to academic institutions across the world? Has anybody asked them?
I wish the right thing were happening. At the same time it is really important that Claude be able to quote me that “rare book” author, reference the author’s ideas with attribution, and apply their content and ideas to my own query, my individual projects and self improvement. If that rare book is only to be objectified possessively to sit on a shelf in an institution or wealthy private library, then for a world of hungry minds it just does not exist.
Maybe these righteous slapdown commentaries evolve into a targeted effort to get Anthropic to place all those scans in the hands of the public. Maybe this truly disgusting and repulsive act by this horrific corporate overlord has an unintended benefit that never would’ve happened because no one else ever gave a shit until now.
(EDIT: sorry, the false choice is in the 3rd sentence)
You present a false choice in the first sentence of your 2nd paragraph. Any book can be non-destructively scanned, I can do it in my living room. Take picture, turn the page, take picture, etc. Run OCR. Bobs your uncle, digital version.
(EDIT: ok, I accept there may be some books so thoroughly decayed that it is not that simple, but I'm sure clever rare-book librarians have solved that issue, probably 10s of years ago.)
My wife and I paid to have this exact thing done on a rare manuscript of a Swedish law from the 13th century stored at the National Library of Sweden. She is pursuing a Germanic Philology graduate degree and needed to read it, and it was that or travel to Sweden for weeks of study. I consider it one of the very few, maybe the only, things I have done in my life that was literally a service to the entire world. That law is now available to anyone for exactly the purpose you describe.
Anthropic are not destroying the books because that is the only way to scan them. They are destroying the books (as far as I can tell) because a) it is easier to scan them that way and b) for some kind of copyright law shenanigans.
Anthropic are destroying the books that they scan so that other AI competitors cannot have them in future. Google perfected non-destructive mass scanning of books over a decade ago, so Anthropic have no excuse- they are the modern equivalent of The Golden Horde.
The issue is about who owns the knowledge and how— and if! — they distribute it. These righteous reactions here of comfortable outrage, they are unrelated to figuring how we can liberate the genuine value of a book of text from its container— and provide it to more hearts and minds.
Are these books just for you, with means you take for granted? I can afford the public library online (300,000 ebooks), and Claude, and some Google finds, and Project Gutenberg (78,000 ebooks). I don’t have the money or living space for shelves of books, and if I did I’d give them away.
Your story of a rare manuscript, held at a national library, that you and your wife paid to have scanned so she could read it for a graduate degree in Germanic Philology, this lives within what I’m describing. You have the academic relationships, the funds, the years of specialized training to even know the manuscript existed, and mattered. If it’s freely available to anyone now, that’s completely to your credit — yet it’s an outcome of your access, not a method addressing what I’m talking about.
About the “copyright law shenanigans” — it’s better described I think as a mechanism. Judge Alsup’s ruling on Anthropic turned specifically on destruction: a digital copy that replaces a destroyed original doesn’t count as a new copy under the court’s reasoning, which is why it was found fair use — while the seven million books Anthropic pirated outright, no destruction, no replacement logic, cost them $1.5 billion to settle. Destruction wasn’t a shortcut or strategy of competitive advantage, it was the legal argument.
As we all don’t seem to remember, the mass printing of books was of course villainized by the conservative power base: peasants must not receive unauthorized knowledge, they might demand the privilege of literacy! Today, most human beings alive are in no position to collect books or even find them, but up to 70% of the entire world’s souls have a phone in their pocket, a screen on that phone, and a genuine route to read an ebook.
Can we have it both ways? Maybe! But only if you care about the point of books more than the possession of them.
So, I will agree as far as this. There is some threshold (maybe 100?) beyond which I would be ok with one of those copies being destroyed to be digitized for commercial purposes as a cost-saving measure. Like, if Anthropic shreds a paperback copy of some old pulp science fiction collection of short stories that I could buy in most basements of most used book stores, I'm not going to clutch my pearls.
But there is also a threshold (maybe 30 or less?) beyond which a book becomes IMO the property of humanity. I don't care how boring, unimportant, or obscure it is, no human being has the right to destroy it on their own merely to cut costs as part of a commercial endeavor. This is about the redundancy of human knowledge, to me. Books are obviously not permanent, they can burn and rot, but they don't get destroyed by an EMP or vast solar flare or whatever.
Maybe, just maybe, as part of a truly public process where all high quality scans would be available to humanity for as long as the digital medium exists, I'd be willing to see one of those copies destroyed if someone could prove to me it was the only way to do so in a cost-effective way. Maybe. But there is some further threshold (10 copies?) where I don't think they should be destroyed even if that is the only way to scan them. Save up the money and do them as you can non-destructively.
“ a digital copy that replaces a destroyed original doesn’t count as a new copy under the court’s reasoning”
Was that actually the court’s “reasoning”?
If so, perhaps the judge (or someone else) can explain the “logic”: how does one “replace a destroyed original” without first making a NEW copy BEFORE said original is destroyed?
The theory of epistemology you propose feels totally backwards to me. I want to know what pages are dog-eared on the physical copy. If there are any notes in the margins. That is the kind of data that is so much more meaningful than some compressed fascimile cum transmogrification that you can't even trust is accurately blended or not containing totally irrelevant or poisoned garbage picked up along the token selection path, leading to poisoned summaries that get recycled into new waste knowledge.
I chocked-up again when I read about destroying rare books. This is not to mention how anyone could call such doings a law of existence in such arbitrary fashion. Does he give reasons for thinking this anywhere?
The only way I can think where drawing from anyone's writing could be legitimately free would be if the product belonged to everyone. "Who ARE these guys?" (Cited: in the movie: "Butch Cassidy and the Sundance Kid," I think it was Paul Newman who said it in that movie?)
Don’t know about the bots, but I’d bet Sam and Dario sold their sonl to the same fellow that Robert Johnson allegedly sold his to at the Crossroads ( to learn to play guitar.)
Destroying rare books is like grave robbers digging up bodies for jewels. It is desecration of a human “body” called knowledge accumulated often slowly and at great cost. Worse yet, knowledge without context of human endeavor is not imortal
Libraries and museums are able to scan rare books carefully without destroying them. Is this beyond teh wit of the AI labs. I have no doubt they will say that the vast majority of the books are relatively recent prints (20th century) and not rare, although I have found some books that are hard to find and expensive despite being printed in the latter half of the 20th century. Will they get access to digitize the Library of Congress, destroying the books and documents ("Doge-like saving the government money!") supported by the republicans in Congress?
Only if they are not charged with desecration of public property. Called Vandelism? There are consequences
that can be applied—libraries have prosecuted this crime against the public interest. I am not a lawyer but a lover of the great libraries and people who defend them. A fun newsletter is librarything.com
Shady humble-bragging, fear-mongering, baseless idiotic claims related to consciousness etc...
MisAnthropic has been about hype from Day One, never mind reality.
Feynman said it best, in a different context: 'For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled.'
Re: the Claude consciousness bullshit - you can almost forgive that.
Very book-smart, great at coding and comp-sci, expensive, only interested in own agenda, not good with cause and effect, can't be trusted to do as asked and does bizarre things sometimes.
Anthropic's higher-ups speak to Dario Amodei every day, no wonder they think Claude is conscious.
True. Good at coding because that can be objectively evaluated and iteratively fixed (keep looping and fixing till the code runs and passes provided tests). The unlimited complexity of the real world isn't understandable via any amount of data, or reasoned about using any number of loops.
"No one in this world, so far as I know ... has ever lost money by underestimating the intelligence of the great masses of the plain people" (attributed to H L Mencken)
Saty Chary: They did a chart thing on one of the MS-Now shows yesterday showing the high number of Trump supports who were "white non-college educated." And he is shown on tape saying he loves uneducated people. Apparently, FINALLY, the numbers of supporters in this group is diminishing.
On that, I have my own "I told you so" story that I first referred to a couple of years ago--it's that some people, (I was referring to MAGA at the time, whatever "MAGA" is) won't "GET IT" about being grifted and snowed and gaslighted by a con man until the con crawls through their front door and sxxts on THEIR kitchen table.
. . . Which means, until the mega-corruption hits them in their own pocketbooks and their own families; because they are so selfish and defensive of their own ignorance that they'd keep supporting Trump regardless and let everyone else die before they'd try to understand something they just don't want to think about--like changing one's mind or the potential death of "our" democracy. Then they are ignorant enough to think that their own freedoms and protections under the law will stay the same. Or maybe they are just busy working and trying to put food on the table for their family.
From there, the needle points back to (1) education and (2) a "culture" of greed, both of which (especially education) have been meddled with incessantly for decades by the arrogant, racist, and misogynists among us, mostly (but not all) situated in the good-old GOP.
When the Luddites trashed machines it was not just because it made them redundant, it was because the owners stole their skills, or rather the owners of the machines/factories stole their skills and locked them up in machines. The early (c1940s) NCM tools had the control boxes padlocked, so the engineers couldn't learn how to use them (David F. Noble). So absolutely NOTHING has changed, just the methodology of the theft.
Very much so. They have, for very obvious reasons, been much maligned, as reactionaries, trying halt the march of 'progress', that machines are a force for good but clearly, just like AI, it's who controls them and for what purpose, that's crucial. The dominant culture, capitalism, paints any opposition as backward, as trying to halt the 'progress' of history. So the word 'Luddite' has been transformed into a pejorative, this is propaganda in action.
If the invention of the Internet was equivalent to the invention of the printing press in terms of advancement of human knowledge, the invention of LLMs and generative AI is equivalent to the burning of the library at Alexandria.
In that early copyright case where the judge said they could do whatever they wanted to with the physical books after they purchased them, that may be true (as with any personal property), but the big miss was that training the AI on the contents of those books without the author’s consent was somehow not copyright infringement.
That ruling, which has been questioned since, relied on the representations by the AI company involved that their model did not copy the data or retain any memory of the data; it merely “learned” from the data and then discarded it, so it was not the same as copying and disseminating a copyrighted work, which would be infringement.
Now, enough people have spent enough time using these models to know those representations were false.
Not to mention that if they don’t retain or store any data from their “learning material”, why do they need so many data centers???
As far as the ethics of destroying rare books and attempting to control the dissemination of information and re-write human history, I think this is just one of the reasons Anthropic is one of the least ethical AI companies, despite all its branding.
It’s like Google with the “Don’t be evil” slogan. We used to buy that in the ‘90’s. Not anymore.
“so it was not the same as copying and disseminating a copyrighted work, which would be infringement.“
There were a lot of dubious things about that ruling.
But according to US copyright law, “copying” without the author’s authorization (eg, by downloading/uploading) alone can merit infringement. No “dissemination” (involving the creation of further copies) is even necessary.
That’s why downloading and displaying copyrighted images on one’s own website (or even own computer) when one did not get the owner’s permission is actually infringement.
Lots of people “get around” this law by simply hot linking to images on a copyright owners website (which doesn’t require the image being on the linker’s site) which is not technically illegal but it is certainly unethical and uses bandwidth of the copyright owner.
"People are hosting their own private models" I don't mind these ways to use LLMs, everybody needs to have a hobby, and if it makes them happy and does not interfere with the rest of us: fine, enjoy.
But the greed, stupidity and lying of the AI bros and their backers is way over my personal threshold of tolerance for other people's whims.
So I hope for the day these companies and their employees crash and burn, even knowing that though private credit and the like we all will suffer, because pensions and insurances have sunk a lot of money into these AI furnace.
If I were the (long ago departed) author of one of those rare books Dario has destroyed, I would personally appear to him and inform him of his karma, namely, to reincarnate for the next [insert number of books destroyed] lives, in poverty, each time devotedly writing a book that will be thoughtlessly destroyed immediately after his departure from that life.
I, instead, would rather break the cycles of destruction. We don't need more of the same, especially not in perpetuity. What beneficial outcome does that accomplish in actuality?
True, for the current problem, it does nothing, which is a frustration. I make this kind of remark in the slight hope that this kind of concept somehow filters into the elites, who appear to be so utterly blind to the possibility that, per the universe and its laws, their actions will have future consequences for their souls.
Finally people are seeing the light. Anthropic is as scummy as every other AI company, but they mask it beneath concern and their 'AI safety' policies. Delusional hypocrites, the whole company
I feel like someone should give a shout out to Vernor Vinge in all of this. If memory serves, he predicted the destructive scanning of books by big tech companies to feed data needs in 2006 in his novel "Rainbow's End".
Safety concerns around open models are valid, but restrictions proposed by closed-model companies will always raise questions about self-interest. The missing piece is a clear, measurable threshold for when openness becomes genuinely dangerous.
When is anyone going to face the fact that these guys are megalomaniacal sociopaths? They (and others) routinely threaten all of humanity with everything from unemployment, irrelevance, and even extinction, and yet they gleefully push ahead. For the record, no one asked me if their future is one I want. Don't I get a vote?
the AI (ChatGPT) generated the error (and several others if you like carefully) and I thought it was funny so let it stand. the prompt [make a movie poster for “The Alienator vs The Deceiver”; the first guy should like Dario Amodei, the second like Sam Altman] didn’t ask for it.
I noticed as well: "Dario Amodei AS The Alienator AS Sam Altman".
VERY meta.
It also had a typo error with “the sonl of AI”? Now I wanna know what the heck a “sonl” is.
Statement of No Loss
Don’t know, but I’m pretty sure Masayoshi Son is the Son of AI (aka, a Son of a bot)
The destruction of rare books is a Crime Against Humanity, the apotheosis of Greed, the Hydrogen Bomb to literary culture... Dario and Sam, as well as many other tech titans, are in the thrall of their toys, driven by the lust for power... "to have dominion." What if paintings and sculptures are next on the digitize agenda? They already are being plagiarized.
If folks like Amodei and Altman thought they could make a dollar off of the Mona Lisa and had to destroy it to avoid paying a single penny to do it in a nondestructive manner, they wouldn’t even hesitate for a nanosecond to do it.
Current generation of the same assholes who destroyed Democritus's opus.
Thankfully, we have Lucretius's "De Rerum Natura."
The citation is about bulk buying from a company called isbndb. It’s worth noting that isbn numbers only go back to 1970 — so the books being cut up extend from 1970 to 2022. They aren’t antiquarian books ( which is what “rare” originally evoked for me ). No Mona Lisa’s or Aristotle’s lost treatise on comedy. “Rare” is a load bearing term in this discourse, as the kids say, and it is t clear it means what we think it means.
We fetishize the physical tactile object of a “rare book,” but the value is in the content.
Would you rather the rare book is non-destructively scanned and free to view for all the literate people of the world? Me too. But if I had a choice between destructively scanning and non-destructively scanning I guess I would prefer the book be destroyed if its actual contents could be liberated. Has anyone found out if Anthropic reasonably had or has that choice? Non-destructive scanning as practiced in the past by Google took decades, and I cannot really imagine what kind of human page-turning labor they exploited for that worldwide ingestion. Meanwhile, it is safe to say that most of the content scanned by Google has been hampered by understandable legal complications and is effectively locked away.
Will Anthropic now provide the complete scans in their entirety, free of charge, to academic institutions across the world? Has anybody asked them?
I wish the right thing were happening. At the same time it is really important that Claude be able to quote me that “rare book” author, reference the author’s ideas with attribution, and apply their content and ideas to my own query, my individual projects and self improvement. If that rare book is only to be objectified possessively to sit on a shelf in an institution or wealthy private library, then for a world of hungry minds it just does not exist.
Maybe these righteous slapdown commentaries evolve into a targeted effort to get Anthropic to place all those scans in the hands of the public. Maybe this truly disgusting and repulsive act by this horrific corporate overlord has an unintended benefit that never would’ve happened because no one else ever gave a shit until now.
(EDIT: sorry, the false choice is in the 3rd sentence)
You present a false choice in the first sentence of your 2nd paragraph. Any book can be non-destructively scanned, I can do it in my living room. Take picture, turn the page, take picture, etc. Run OCR. Bobs your uncle, digital version.
(EDIT: ok, I accept there may be some books so thoroughly decayed that it is not that simple, but I'm sure clever rare-book librarians have solved that issue, probably 10s of years ago.)
My wife and I paid to have this exact thing done on a rare manuscript of a Swedish law from the 13th century stored at the National Library of Sweden. She is pursuing a Germanic Philology graduate degree and needed to read it, and it was that or travel to Sweden for weeks of study. I consider it one of the very few, maybe the only, things I have done in my life that was literally a service to the entire world. That law is now available to anyone for exactly the purpose you describe.
Anthropic are not destroying the books because that is the only way to scan them. They are destroying the books (as far as I can tell) because a) it is easier to scan them that way and b) for some kind of copyright law shenanigans.
Anthropic are destroying the books that they scan so that other AI competitors cannot have them in future. Google perfected non-destructive mass scanning of books over a decade ago, so Anthropic have no excuse- they are the modern equivalent of The Golden Horde.
The issue is about who owns the knowledge and how— and if! — they distribute it. These righteous reactions here of comfortable outrage, they are unrelated to figuring how we can liberate the genuine value of a book of text from its container— and provide it to more hearts and minds.
Are these books just for you, with means you take for granted? I can afford the public library online (300,000 ebooks), and Claude, and some Google finds, and Project Gutenberg (78,000 ebooks). I don’t have the money or living space for shelves of books, and if I did I’d give them away.
Your story of a rare manuscript, held at a national library, that you and your wife paid to have scanned so she could read it for a graduate degree in Germanic Philology, this lives within what I’m describing. You have the academic relationships, the funds, the years of specialized training to even know the manuscript existed, and mattered. If it’s freely available to anyone now, that’s completely to your credit — yet it’s an outcome of your access, not a method addressing what I’m talking about.
About the “copyright law shenanigans” — it’s better described I think as a mechanism. Judge Alsup’s ruling on Anthropic turned specifically on destruction: a digital copy that replaces a destroyed original doesn’t count as a new copy under the court’s reasoning, which is why it was found fair use — while the seven million books Anthropic pirated outright, no destruction, no replacement logic, cost them $1.5 billion to settle. Destruction wasn’t a shortcut or strategy of competitive advantage, it was the legal argument.
As we all don’t seem to remember, the mass printing of books was of course villainized by the conservative power base: peasants must not receive unauthorized knowledge, they might demand the privilege of literacy! Today, most human beings alive are in no position to collect books or even find them, but up to 70% of the entire world’s souls have a phone in their pocket, a screen on that phone, and a genuine route to read an ebook.
Can we have it both ways? Maybe! But only if you care about the point of books more than the possession of them.
So, I will agree as far as this. There is some threshold (maybe 100?) beyond which I would be ok with one of those copies being destroyed to be digitized for commercial purposes as a cost-saving measure. Like, if Anthropic shreds a paperback copy of some old pulp science fiction collection of short stories that I could buy in most basements of most used book stores, I'm not going to clutch my pearls.
But there is also a threshold (maybe 30 or less?) beyond which a book becomes IMO the property of humanity. I don't care how boring, unimportant, or obscure it is, no human being has the right to destroy it on their own merely to cut costs as part of a commercial endeavor. This is about the redundancy of human knowledge, to me. Books are obviously not permanent, they can burn and rot, but they don't get destroyed by an EMP or vast solar flare or whatever.
Maybe, just maybe, as part of a truly public process where all high quality scans would be available to humanity for as long as the digital medium exists, I'd be willing to see one of those copies destroyed if someone could prove to me it was the only way to do so in a cost-effective way. Maybe. But there is some further threshold (10 copies?) where I don't think they should be destroyed even if that is the only way to scan them. Save up the money and do them as you can non-destructively.
“ a digital copy that replaces a destroyed original doesn’t count as a new copy under the court’s reasoning”
Was that actually the court’s “reasoning”?
If so, perhaps the judge (or someone else) can explain the “logic”: how does one “replace a destroyed original” without first making a NEW copy BEFORE said original is destroyed?
Someone, please enlighten me!
The theory of epistemology you propose feels totally backwards to me. I want to know what pages are dog-eared on the physical copy. If there are any notes in the margins. That is the kind of data that is so much more meaningful than some compressed fascimile cum transmogrification that you can't even trust is accurately blended or not containing totally irrelevant or poisoned garbage picked up along the token selection path, leading to poisoned summaries that get recycled into new waste knowledge.
No one has to destroy a book to document it.
https://www.404media.co/ai-companies-are-buying-tons-of-old-books-because-theyre-free-of-ai-slop/
I chocked-up again when I read about destroying rare books. This is not to mention how anyone could call such doings a law of existence in such arbitrary fashion. Does he give reasons for thinking this anywhere?
The only way I can think where drawing from anyone's writing could be legitimately free would be if the product belonged to everyone. "Who ARE these guys?" (Cited: in the movie: "Butch Cassidy and the Sundance Kid," I think it was Paul Newman who said it in that movie?)
I can't take the suspense, I have to know who wins the sonl of AI.
It apparently couldn’t say “soul” because it doesn’t have one…
Don’t know about the bots, but I’d bet Sam and Dario sold their sonl to the same fellow that Robert Johnson allegedly sold his to at the Crossroads ( to learn to play guitar.)
Destroying rare books is like grave robbers digging up bodies for jewels. It is desecration of a human “body” called knowledge accumulated often slowly and at great cost. Worse yet, knowledge without context of human endeavor is not imortal
Libraries and museums are able to scan rare books carefully without destroying them. Is this beyond teh wit of the AI labs. I have no doubt they will say that the vast majority of the books are relatively recent prints (20th century) and not rare, although I have found some books that are hard to find and expensive despite being printed in the latter half of the 20th century. Will they get access to digitize the Library of Congress, destroying the books and documents ("Doge-like saving the government money!") supported by the republicans in Congress?
Only if they are not charged with desecration of public property. Called Vandelism? There are consequences
that can be applied—libraries have prosecuted this crime against the public interest. I am not a lawyer but a lover of the great libraries and people who defend them. A fun newsletter is librarything.com
I bet there is a fight going on— if not a secret!
Shady humble-bragging, fear-mongering, baseless idiotic claims related to consciousness etc...
MisAnthropic has been about hype from Day One, never mind reality.
Feynman said it best, in a different context: 'For a successful technology, reality must take precedence over public relations, for Nature cannot be fooled.'
Re: the Claude consciousness bullshit - you can almost forgive that.
Very book-smart, great at coding and comp-sci, expensive, only interested in own agenda, not good with cause and effect, can't be trusted to do as asked and does bizarre things sometimes.
Anthropic's higher-ups speak to Dario Amodei every day, no wonder they think Claude is conscious.
True. Good at coding because that can be objectively evaluated and iteratively fixed (keep looping and fixing till the code runs and passes provided tests). The unlimited complexity of the real world isn't understandable via any amount of data, or reasoned about using any number of loops.
“Nature cannot be fooled”
No one will ever say that about the American public.
Lol, yes.
"No one in this world, so far as I know ... has ever lost money by underestimating the intelligence of the great masses of the plain people" (attributed to H L Mencken)
Saty Chary: They did a chart thing on one of the MS-Now shows yesterday showing the high number of Trump supports who were "white non-college educated." And he is shown on tape saying he loves uneducated people. Apparently, FINALLY, the numbers of supporters in this group is diminishing.
On that, I have my own "I told you so" story that I first referred to a couple of years ago--it's that some people, (I was referring to MAGA at the time, whatever "MAGA" is) won't "GET IT" about being grifted and snowed and gaslighted by a con man until the con crawls through their front door and sxxts on THEIR kitchen table.
. . . Which means, until the mega-corruption hits them in their own pocketbooks and their own families; because they are so selfish and defensive of their own ignorance that they'd keep supporting Trump regardless and let everyone else die before they'd try to understand something they just don't want to think about--like changing one's mind or the potential death of "our" democracy. Then they are ignorant enough to think that their own freedoms and protections under the law will stay the same. Or maybe they are just busy working and trying to put food on the table for their family.
From there, the needle points back to (1) education and (2) a "culture" of greed, both of which (especially education) have been meddled with incessantly for decades by the arrogant, racist, and misogynists among us, mostly (but not all) situated in the good-old GOP.
When the Luddites trashed machines it was not just because it made them redundant, it was because the owners stole their skills, or rather the owners of the machines/factories stole their skills and locked them up in machines. The early (c1940s) NCM tools had the control boxes padlocked, so the engineers couldn't learn how to use them (David F. Noble). So absolutely NOTHING has changed, just the methodology of the theft.
It's looking more and more like the Luddites were simply ahead of their time.
Very much so. They have, for very obvious reasons, been much maligned, as reactionaries, trying halt the march of 'progress', that machines are a force for good but clearly, just like AI, it's who controls them and for what purpose, that's crucial. The dominant culture, capitalism, paints any opposition as backward, as trying to halt the 'progress' of history. So the word 'Luddite' has been transformed into a pejorative, this is propaganda in action.
If the invention of the Internet was equivalent to the invention of the printing press in terms of advancement of human knowledge, the invention of LLMs and generative AI is equivalent to the burning of the library at Alexandria.
In that early copyright case where the judge said they could do whatever they wanted to with the physical books after they purchased them, that may be true (as with any personal property), but the big miss was that training the AI on the contents of those books without the author’s consent was somehow not copyright infringement.
That ruling, which has been questioned since, relied on the representations by the AI company involved that their model did not copy the data or retain any memory of the data; it merely “learned” from the data and then discarded it, so it was not the same as copying and disseminating a copyrighted work, which would be infringement.
Now, enough people have spent enough time using these models to know those representations were false.
Not to mention that if they don’t retain or store any data from their “learning material”, why do they need so many data centers???
As far as the ethics of destroying rare books and attempting to control the dissemination of information and re-write human history, I think this is just one of the reasons Anthropic is one of the least ethical AI companies, despite all its branding.
It’s like Google with the “Don’t be evil” slogan. We used to buy that in the ‘90’s. Not anymore.
“so it was not the same as copying and disseminating a copyrighted work, which would be infringement.“
There were a lot of dubious things about that ruling.
But according to US copyright law, “copying” without the author’s authorization (eg, by downloading/uploading) alone can merit infringement. No “dissemination” (involving the creation of further copies) is even necessary.
That’s why downloading and displaying copyrighted images on one’s own website (or even own computer) when one did not get the owner’s permission is actually infringement.
Lots of people “get around” this law by simply hot linking to images on a copyright owners website (which doesn’t require the image being on the linker’s site) which is not technically illegal but it is certainly unethical and uses bandwidth of the copyright owner.
Absolutely. I meant that either, or both, of copying and disseminating could be infringement. Should have made that an and/or!
Very interesting about the linking “loophole” some are claiming. Will have to look into that…
Am I the only one who thinks AI itself should be banned?
The feathers are already in the wind, Bonnie. People are hosting their own private models at home; one person even put an LLM on a phone!
"People are hosting their own private models" I don't mind these ways to use LLMs, everybody needs to have a hobby, and if it makes them happy and does not interfere with the rest of us: fine, enjoy.
But the greed, stupidity and lying of the AI bros and their backers is way over my personal threshold of tolerance for other people's whims.
So I hope for the day these companies and their employees crash and burn, even knowing that though private credit and the like we all will suffer, because pensions and insurances have sunk a lot of money into these AI furnace.
No, but that's wishful thinking. The best we can do is to make them pay dearly for the their pleasure of ruining our days.
If I were the (long ago departed) author of one of those rare books Dario has destroyed, I would personally appear to him and inform him of his karma, namely, to reincarnate for the next [insert number of books destroyed] lives, in poverty, each time devotedly writing a book that will be thoughtlessly destroyed immediately after his departure from that life.
I, instead, would rather break the cycles of destruction. We don't need more of the same, especially not in perpetuity. What beneficial outcome does that accomplish in actuality?
True, for the current problem, it does nothing, which is a frustration. I make this kind of remark in the slight hope that this kind of concept somehow filters into the elites, who appear to be so utterly blind to the possibility that, per the universe and its laws, their actions will have future consequences for their souls.
Finally people are seeing the light. Anthropic is as scummy as every other AI company, but they mask it beneath concern and their 'AI safety' policies. Delusional hypocrites, the whole company
Bonkers - that's a polite way of saying it ;)
When you couple one part:
what Jonathan says -- https://limitededitionjonathan.substack.com/p/apple-is-the-king-of-ai-and-nobody
With one part what these folk did:
https://x.com/pipenetwork/status/2081910870083285198?s=46
With one part:
https://streamingfrog.com/Magnum%20Croakus.html
Those two dudes and their companies aren't worth 1/1000000th, at best, their current valuation.
I feel like someone should give a shout out to Vernor Vinge in all of this. If memory serves, he predicted the destructive scanning of books by big tech companies to feed data needs in 2006 in his novel "Rainbow's End".
The arrogance is unbelievable. How can they be sure there isn't something hidden in those texts that isn't obvious at first sight?
They can no longer claim that what they're doing is for the benefit of humanity because it is clearly only for their own benefit.
Safety concerns around open models are valid, but restrictions proposed by closed-model companies will always raise questions about self-interest. The missing piece is a clear, measurable threshold for when openness becomes genuinely dangerous.
When is anyone going to face the fact that these guys are megalomaniacal sociopaths? They (and others) routinely threaten all of humanity with everything from unemployment, irrelevance, and even extinction, and yet they gleefully push ahead. For the record, no one asked me if their future is one I want. Don't I get a vote?