Did you realize that we live in a reality where SciHub is illegal, and OpenAI is not?
Did you realize that we live in a reality where SciHub is illegal, and OpenAI is not? 110 comments
@dadadan @yabellini This site literally saved me during my university time as a political science student @yabellini to be fair, sampling in music is without permission is also illegal, but if you distort the sample enough to be hard to algorithmically detect, you'll probably get away with it. I guess the counterpoint is that scihub is stealing from the rich and giving to the poor, while OpenAI is doing the opposite. @kinyutaka @yabellini that's my point. IMO it's an argument in favour of LLMs. Personally I don't think copyright should exist at all. But while it does, the fact that you can get away with breaching it if you are rich enough sucks. I think copyright should exist. For 28 years, like it used to... If you can't profit off your work after a quarter of your life, maybe it's time for other hands to give it a try. @kinyutaka @yabellini that would definitely be a good first step! Also non-transferability (no selling or inheriting) Though, really, if the limit is 28 years, you can make the argument for inheritance, because the family didn't get the benefit of the full earnings. The fact that I could draw a new character for a one off story and it is copyrighted for decades after I die is horseshit. Popeye is older than my dead grandparents and we still have to wait to draw pictures of him. @kinyutaka @naught101 @yabellini Screw inheritance. Give all children access to free education, healthcare and sufficient funds for a healthy (not poverty stressed) upbringing. Let's not have them own trusts and shares and real estate based on their parents'/ancestors' exploitation of others. Also. Let's take away the 'ability' of legal bodies to own anything. @grumble209 @kinyutaka @naught101 @yabellini Non-native, I probably picked the wrong word. Legal persons, legal entities? @zombiecide @grumble209 @naught101 @yabellini So, no inheritance for anything? Not homes, not intellectual property, not favored possessions? What happens to the things that would otherwise be inherited? @kinyutaka @zombiecide @grumble209 @naught101 @yabellini Inheritance should only be for personal possessions. Furniture, mementos, etc. @adaliabooks @kinyutaka @grumble209 @naught101 @yabellini My take on that is "Do you need two chairs? No, your arse only covers one? Then you only get one chair.*" I'm all for everyone getting some kind of voucher for basic housing on reaching adulthood, which could be put into buying part of your parents' place (the rest you'd have to pay from wages), or in a coop so you can move more easily etc. Land should be in the commons + community administered *big chairs, wheelchairs are available @zombiecide @adaliabooks @grumble209 @naught101 @yabellini As someone who doesn't have a house, I look forward to getting what little cash I get from an inheritance, which would be enough to get a down payment going on my own. I believe that every adult should have the means to own a home, but home ownership implies being able to keep the home and pass it to children. Where you have a point is the people with the million dollar mansion and 5 vacation homes. @kinyutaka "home ownership implies being able to keep the home and pass it to children" My idea is based on trying to balance needs, fairness, realities of life and environmental impact. In my country, per capita living space has been growing (almost 50m2 for a single person atm) while about 10% live in cramped quarters (more than one person per room, say around 10m2 or less per person), thanks to policy based on self-owned single homes and apartment blocks for rent being supported 1/2 @kinyutaka Many apartment blocks end up in the hands of shady companies located in some tax haven that don't do the legally required upkeep, while in detached homes, adult children move out, people become elderly, they have to pay to maintain the space, they're reliant on a car-based surburban infrastructure even if too frail for a car, their house can't be retrofitted for a wheelchair but they'd lose the only thing their children can inherit, so they stay as long as possible. 2/2 Oh, believe you me, I am not a huge fan of apartment living, but that's what will end up happening if we don't allow and encourage individual home ownership. Here is a proposal, rental inheritance. If a house is being rented out, then the renter inherits the house upon the death of the owner, instead of the children. And a massive penalty to transferring ownership of a rental home to anyone but the renter, including as a gift or a sale. Like, if I am living in a home worth $250,000, and the owner passes the home to his kid to prevent it from going to me, it could cost $225,000 in extra taxes. Congrats, you can keep the house, but you'll never profit. Or they could sell the home to me early, and even if I only give them $100,000 it's more than they'd get from keeping the home and renting it more. And if they just wait, I get the home free. @kinyutaka You just ignored the issued I mentioned about suburbanization, families growing, shrinking and members getting elderly. And with the same floor space in a detached/semi house requiring more resources, land and energy than in an apartment block. Coop housing can cope with all that, syndicated housing even better. I am not sure how co-op or syndicated housing is supposed to work, so I can't speak on that. @kinyutaka @zombiecide @grumble209 @naught101 @yabellini But the point is that the reason you don't have a home is because of the shitty capitalist system, of which inheritance is a part. @adaliabooks @zombiecide @grumble209 @naught101 @yabellini I guarantee you that if the government gives everyone a home, it's going to be a glorified apartment, with even less maintenance. @kinyutaka @zombiecide @adaliabooks @grumble209 @yabellini The problem is property investors. @adaliabooks @kinyutaka @zombiecide @grumble209 @naught101 @yabellini this is a very tricky and delicate topic. We should be humble and accept any real solution would be complex and should ensure social justice AND individual liberties, wich sounds beautiful but hard to achieve. I wish some things were different but I only know it would be hard to do the right thing for everyone. @naught101 @kinyutaka @yabellini So ideally, scientists publishing in journals would retain copyright on their work, and give the journal a non-exclusive license to publish. @svgeesus @rochelimit @naught101 @kinyutaka @yabellini And some of those journals make authors pay for open publishing. @svgeesus @naught101 @kinyutaka @yabellini yes, copyright should exist, but with an actual goal of supporting creators. The way copyright is *used* now makes no sense. But neither does removing it entirely. @naught101 @yabellini openai is MUCH WORSE than sampling. Every pixel, every word their tools generate, has been generated from stolen content. Meanwhile few music is actually 100% made of stolen samples. @yabellini Huh, I'd never actually considered this and it kinda blew my mind. Reproducing copyrighted material is only bad when you do it for personal gain and not corporate profits, I suppose. Don't like that! 🤔 "Training" a "neural net" that is connected in such a way as to always reproduce its input exactly. Make "SciHub AI", which will generate documents from a machine trained in this way. 🤔 Yes, it is quite astonishing! SciHub, a platform that offers free access to scientific research articles, is considered illegal, while OpenAI, an organization dedicated to advancing artificial intelligence, operates within legal boundaries. It is remarkable how our society often prioritizes restricting knowledge instead of encouraging it. This inconsistency brings up significant concerns about the current state of sharing information and the importance of open access to scientific research. @sharpsolverub It also highlights that #copyright and #patents as they stand today are broken. I think a case could be made that its ideals of protecting and encouraging new ideas was always skewed toward protecting the already wealthy, but this just highlights that we need major reform. @yabellini In my opinion, they should both be legal, provided OpenAI stops profiting from their use of copyrighted material (which they are). @yabellini Awesome, thank you. I was unaware of SciHub. As a writer and former graphic designer, I have always believed in copyright laws. However, with the climate emergency, the 6th extinction and plastics in our bodies, among other things, access to direct, unfiltered, sources of scientific information are absolutely essential. @gdeihl @yabellini your perspective on copyright regarding artistic creations is totally valid. But science is almost always publicly funded, and often produced by sampling from the public - biologically and otherwise. In a very important way, it belongs to the people. And even most authors of scientific work oppose paywalling it. @Gothilla @yabellini Good points. Scientific discoveries are happening at light speed now as well. The average person should have a voice in how these discoveries are used. Science is cool, how it's implemented, not always so. @yabellini They are not defending copies right, they are defending capital. The law are made for reinforcement of the richest, that's all. There have been serious clues this was where cyberspace was headed. After MIT helped kill Aaron Swartz for liberating JSTOR content for use beyond Ivory towers, they jailed Jeremy Hammond for exposing GIfiles then set about slow motion execution of Julian Assange while Palintir & Google grow their Federal contract portfolios w surveillance technologies.. hints abound! https://wikileaks.org/gifiles/docs/29/2956566_palantir-the-war-on-terror-s-secret-weapon-.html @yabellini because law is for the legal experts and the rich to protect from the masses. it doesnt work and was never intended for the masses. that is just the propaganda to keep the masses quiet. @yabellini SciHub makes papers public that are behind paywalls. I agree, that they shouldn't be behind paywalls, but it's completely different to OpenAI. I recommend reading the lawsuit, it was not only written by lawyers who know the law but it is also very clear: https://nytco-assets.nytimes.com/2023/12/NYT_Complaint_Dec2023.pdf @yabellini I can not read the article as it's behind a paywall and the other document is 69 pages long. I will not read that. If you want to say something with it, say waht you want to say. Depending on what you will say, I will think about if I want to check that with the provided sources or not. @skylarkingmullet @duco And here is when they also say it’s “impossible” to create useful AI models without copyrighted material" @yabellini @skylarkingmullet at least under German law, the author of any text has the rights on it. Every Wikipedia article has authors with copyright on it. But they licenced it under a free licence, so everyone can use it. So as every text written by a human, every photo taken by a human and every image painted by a human is copyrighted, OpenAI is correct, that they can not train the AI without that. That does not mean, that the texts are behind a paywall. @duco @yabellini @skylarkingmullet That's inaccurate. When you write on wikipedia you release any rights, you write under CC00, wikimedia however has CC 4.0 BY-SA on all wikipedia content. @skylarkingmullet @yabellini so they sued OpenAI. Well people sued government for legislation of masks against Corona. Just because someone sues someone doesn't mean they are right. Let's wait for what the judges say. The second part seems to be about data protection, not copyright. @yabellini they’re both illegal, but OpenAI stealing from commoners and is protected by billionaires while Sci-hub is stealing from billionaires and giving to commoners. @yabellini Perverse as the situation may seem, I appreciate hearing someone who understands this. @yabellini -- Truth. I thought Elsevier put the kabosh on institutions that were uploading papers to SciHub. You can only really access older publications on there now. Anyone found a p2p alternative to SciHub for sharing papers? Or is someone writing one? #openaccess @yabellini so we can make scihub legal by turning it into a public LLM service :thonking:
@yabellini Did you realize that we live in a reality where uploading full movies to youtube is illegal, and collage art is not? tis all bout teh meta data, @yabellini, strip em off sci hub data base, and extract plain text from teh portable documents then concatenate them behind a full text search and no one will dare to question its legality ever again #lyfhax not to mention it’ll be more performant and scalable than what ever open al be cooking these day @yabellini 😲 we'll get there.... Sooner than later, I should hope -- the world's just not ready for it @yabellini I'll probably regret this, but... one is using information posted publicly to the web in what's arguably a transformative fair use. One uses stolen credentials to provide verbatim copies of content with zero pretense of fair use. Yes, it should all be free and open, but wishes don't pay the salaries of all the people and systems required to provide the services. @williamgunn Research is pretty much the paradigmatic example of fair use Also, “the people and systems required to provide the services” no doubt includes the authors of the papers and their research equipment and facilities, which receive $0 from the scandalous fees the publishers charge @magitweeter @williamgunn @yabellini yeah this. Very often, this is even publicly funded research that the publishers just charge ridiculous amounts for, while all they do is distribute a few megabytes of data. Neither the authors nor the peer reviewers get paid anything. Side note: assuming that https://sci-hub.live/ is the actual scihub website, I just visited and my anti-virus software reported a phishing threat. In the past, this sometimes is a false positive. I've sent email with the details to the sci-hub contact email. @yabellini I'm not familiar with this acronym yet. @kranzkrone the search engine of your choice will help you. And I bet so will Wikipedia. Some might compare it with "the PirateBay for scientific literature". There, as expected: https://en.wikipedia.org/wiki/Sci-Hub @yabellini @yabellini OpenAI is almost certainly illegal if you actually read and apply the law as written and intended, while scihub almost certainly isn't. @yabellini It's associated tools are an enabler though for people who lack resources to gain or capture knowledge. Say for a young student aspiring to clear doubts and unable to finance any tutoring. This reminds me back when a PC software maker (Norton, I think) kept calling the free and very useful 'hacker tool' utility Back Orifice 2000 illegal and bad but offered the exact same functionality in a software package *they* sold for money. Big companies think something can only be 'legal' or 'good' or 'acceptable' or 'reliable' if they can make money from it. What I find difficult to believe is that people exist who still hadn't noticed that this is exactly how capitalism works. @yabellini That’s the same reality in which (national) libraries face significant legal challenges to try and selectively capture web resources to digitally preserve them for posterity, while OpenAI industrially hoovers up the web to enable the creation of a massive information oil spill. That might depend on which country you live in. In how many countries has the legality been tested so far?
@yabellini Rich people and big corporations get away with too much illegal stuff. @yabellini it really saddens me that the country I live in is the reason why SciHub is illegal :( I literally just had to ask someone for a paper I'm an author on because my work doesn't have access. @yabellini silly idea: someone should setup a web app that pretends to be an LLM chat bot but it answers your questions with the full text of scientific papers, and writes them gradually similar to Chat GPT responses. @yabellini It's so weird that we live in a reality where copyright prevents the free spread of information and ideas for more than 90 years. @yabellini@fosstodon.org |
@yabellini I absolutely been preaching this out over and over at my University