SoylentNews is people
SoylentNews
Sci-Hub Pledges Open Source & AI Alongside Crypto Donation Drive
Sci-Hub founder Alexandra Elbakyan has launched a donation drive to ensure the operations and development of the popular academic research platform. For safety reasons, donations can only be made in cryptocurrencies but the pledges include a drive to open source the project and the introduction of artificial intelligence to discover new hypotheses.
[...] A new campaign launched by Elbakyan on Saturday hopes to encourage people to contribute to the site's future, promising "dramatic improvements" over the next few years in return.
In addition to offering enhanced search features and a mobile app, Sci-Hub is pledging developments that include the open sourcing of the project. Also of interest is the pledge to introduce an artificial intelligence component that should make better use of the masses of knowledge hosted by Sci-Hub.
"Sci-Hub engine will [be] powered by artificial intelligence. Neural Networks will read scientific texts, extract ideas and make inferences and discover new hypotheses," Elbakyan reveals.
The overall goal of the next few years is to boost content availability too, expanding from hosting "the majority of research articles" available today to include "any scientific document ever published."
Related: Sci-Hub Bounces from TLD to TLD
Sci-Hub Proves That Piracy Can be Dangerously Useful
Paywall: A Documentary About the Movement for Open-Access Science Publishing
Swedish ISP Punishes Elsevier for Forcing It to Block Sci-Hub by Also Blocking Elsevier
Library Genesis Seeding Project Helps to Decentralize Archive of Scientific Knowledge
Scientists to be Heard in High-Profile Publisher Lawsuit Against Sci-Hub in India
(Score: 5, Interesting) by sgleysti on Sunday August 15 2021, @08:11PM (5 children)
If you click the link and read the Q&A with Alexandra, she explains that donations are in cryptocurrency because paypal kept freezing accounts she would use for donations. The only other buzzwordy thing in the list is AI, regarding which I remain skeptical.
The other improvements seem reasonable, modulo some strong wording:
Disclaimer: Alexandra Elbakyan is the only person I'd list as a personal hero, and I have donated to Sci-Hub via bitcoin.
Peeved rant: I have a background in numerical computing, with a focus in nonlinear optimization, and it pains me that a lot of job descriptions in this field only talk about AI and machine learning now. The buzzwords have taken over. Standard methods are still incredibly powerful and useful when appropriately applied, and they continue to have wide application. I regard modern AI as statistical classifiers "tuned" by large data sets.
(Score: 2) by takyon on Sunday August 15 2021, @08:32PM
If the "AI" features do materialize, Sci-Hub could potentially slam something like Meta [wikipedia.org] into the ground. Although "discovering new hypotheses" goes far beyond what Meta tries to do.
[SIG] 10/28/2017: Soylent Upgrade v14 [soylentnews.org]
(Score: 4, Insightful) by legont on Sunday August 15 2021, @08:58PM (1 child)
She is my hero too. I worked in publishing when she alone have done what our whole nonprofit bloody rich joint could not.
"Wealth is the relentless enemy of understanding" - John Kenneth Galbraith.
(Score: 0) by Anonymous Coward on Sunday August 15 2021, @10:15PM
Kazakhstan's second greatest export after Borat?
(Score: 4, Insightful) by bzipitidoo on Monday August 16 2021, @12:58AM
I list Aaron Swartz as another hero.
What I find so infuriating is that IP law is supposed to work for the little person, but it has successfully been turned on its head. For decades, IP law has been the chief enabler of knowledge hoarding and access tolls. It's absolutely ridiculous that an academic publisher asks $30 for a 10 page scientific article, and passes on zero of that to the authors. Additional insult is heaped upon us by their weak justifications that are grossly exaggerated when not flat wrong. They do not help fund the research. They don't review the research either, that work gets farmed out to more unpaid volunteers. The editors and organizers too are likely to be unpaid volunteers, doing the work as part of their job in academia. They scarcely even do the librarian work of preserving, indexing, and most of all, making available. So what do they do? Get the material printed, that's about all. And print media is obsolete now. If academic publishers think I want to hand type in some source code in a journal article that is available only in a print edition, to see for myself the results, they are very much mistaken. Just try to get hold of some obscure work published before 1990. The older it is, the more likely it'll be simply unavailable. Lot of stuff from the 1980s might well be computerized, and some things from the 1970s, but older than that, it was pretty much exclusively done up in a typewriter, and won't have been scanned and OCRed yet. People have been working to change that, but progress is slow, greatly hindered by copyright law. What should be possible is downloading a torrent containing every issue that some academic journal ever published. Shouldn't be more than a few gigabytes, taking only a few minutes to download. But no can do. Academic publishers are straight up useless parasites and thieves. They've tried to gaslight the entire scientific community-- they know they're full of crap.
Swartz was the victim of an overzealous prosecutor throwing the book at him. Probably not a good idea, but I seriously wonder if another group should be formed to jailbreak (an actual, physical jailbreaking, not a hack into an iPhone) individuals who end up imprisoned for having done nothing but enable more copying. The law is notorious about being extremely reluctant to free the wrongfully convicted and imprisoned, because they are more concerned about their own appearances than justice. The US is especially prison happy. Considering what happened to Dmitry Sklyarov when he visited the US, I strongly advise Ms. Elbakyan not to visit, until such time as the IP regime is majorly reformed and transformed.
(Score: 0) by Anonymous Coward on Monday August 16 2021, @06:45AM
a couple of points:
why would you need funds to make something open source?
you just add the desired license to the archive, and put it in a public repository somewhere. it takes ten minutes if you've never done it before, but know how to use a forum.
the people behind arXiv have been doing automated sifting through data etc for quite some time, and they also have a lot of articles (most in plain text).
that's a serious research team with quite some brainpower.
and yet they're not claiming to use AI for new scientific discoveries. I specifically asked Paul Ginsparg this, because in his talk he mentioned a posteriori examinations of breakthrough papers (they can be seen as "papers from field A using a lot of terminology from field B, showing previously unknown connections). And he didn't seem like a modest guy.