Just another news site
Raubbau unter dem rechten Bolsonaro in Rekordtempo. Lula da Silva kündigt Schutzmaßnahmen an. Auch ein Ministerium für die indigenen Völker des Landes soll eingerichtet werden.
Viele deutsche Unternehmen wollen verstärkt in den USA investieren. Ein Grund ist: Dort ist Energie billiger zu haben als in Europa.
Warum man sich beim Faulenzen im Büro wie ein Geheimagent fühlen kann und Tauben als Drohnen: unser Überblick über freie, kostenlos verfügbare Inhalte. Von Martin Wolf (Urheberrecht, Film)
Warum man sich beim Faulenzen im Büro wie ein Geheimagent fühlen kann und Tauben als Drohnen: unser Überblick über freie, kostenlos verfügbare Inhalte. Von Martin Wolf (Urheberrecht, Film) Die Rentenpolitiker der Ampelkoalition kündigen ein Wunder an: Das “Grundkonzept zur Aktienrente”. Ein Kommentar mit Hintergrund.
Galactica language model generated convincing text about fact and nonsense alike.
Enlarge / An AI-generated illustration of robots making science. (credit: Ars Technica)
On Tuesday, Meta AI unveiled a demo of Galactica, a large language model designed to "store, combine and reason about scientific knowledge." While intended to accelerate writing scientific literature, adversarial users running tests found it could also generate realistic nonsense. After several days of ethical criticism, Meta took the demo offline, reports MIT Technology Review.
Large language models (LLMs), such as OpenAI's GPT-3, learn to write text by studying millions of examples and understanding the statistical relationships between words. As a result, they can author convincing-sounding documents, but those works can also be riddled with falsehoods and potentially harmful stereotypes. Some critics call LLMs "stochastic parrots" for their ability to convincingly spit out text without understanding its meaning.
Enter Galactica, an LLM aimed at writing scientific literature. Its authors trained Galactica on "a large and curated corpus of humanity’s scientific knowledge," including over 48 million papers, textbooks and lecture notes, scientific websites, and encyclopedias. According to Galactica's paper, Meta AI researchers believed this purported high-quality data would lead to high-quality output.
Barbaren, Naturmenschen, Seelenmenschen und der aufgeklärte “Westen”: Über Konstruktions – und Denkmuster des deutschen Bildes von Russland. Krieg spielt eine zentrale Rolle.
Holmes is expected to appeal.
Enlarge / Elizabeth Holmes (C), founder and former CEO of blood testing and life sciences company Theranos, walks with her mother Noel Holmes and partner Billy Evans into the federal courthouse for her sentencing hearing on November 18, 2022, in San Jose, California. (credit: Getty | Amy Osborne)
Elizabeth Holmes has been sentenced to 11.25 years in federal prison, plus three years of supervision for her conviction in January on four counts of defrauding investors of her failed blood-testing company, Theranos. Restitution in the case will be determined at a later hearing, not yet set.
The sentencing is less than the maximum of 20 years set by federal sentencing guidelines, but still more than the nine-year prison sentence recommended by the probation officer in Holmes' case.
Federal prosecutors had sought 15 years of imprisonment and for Holmes, 38, to pay roughly $804 million in restitution to defrauded investors. Holmes' lawyers, meanwhile, requested just 18 months of house arrest and argued that she has "essentially no assets" and could not pay a nine-figure fine.
After Z-Library was targeted by U.S. law enforcement, a group of anonymous archivists worked around the clock to get a shadow library search engine online. This week, the ‘team’ behind the Pirate Library Mirror launched “Anna’s Archive” which offers a gateway to various book resources. A brazen move, but one where risks and privacy were carefully considered.
From: TF, for the latest news on copyright battles, piracy and more.
When the file-sharing revolution hit the mainstream well over a decade ago, many site operators and users positioned ‘sharing’ as an ideology.
This altruistic belief has started to fade in recent years. For most pirate sites and services, money is the main driver, with piracy as the means to that end.
There are some notable exceptions of course, especially in the publishing industry where free access to knowledge is advocated by Sci-Hub, LibGen, and others. The law is often blind to these motivations but they are fuel for discussion, even among well-respected scholars.
This year a new team entered the piracy arena in the form of the Pirate Library Mirror. As the name suggests, the team behind it is well-aware of the legal status of their operation.
“We deliberately violate the copyright law in most countries. This allows us to do something that legal entities cannot do: making sure books are mirrored far and wide,” they note.
The goal of the team is to preserve as much written material as possible. This included archiving a full copy of Z-Library, which was completed in September. That was just the start, however.
This week, another chapter was added to the saga. “Anna”, who was one of the driving forces behind the Pirate Library, decided to release her own site, simply called “Anna’s Archive“. While this name sounds benign, its impact certainly isn’t.
Anna’s Archive is basically a meta-search engine that can find content from third-party ‘pirate’ sources. This includes direct downloads through Library Genesis forks and Z-Library, two well-known shadow libraries. In addition, torrent links are included as well.
The site presents all information in a slick and easy-to-navigate format that rivals commercial vendors. However, Anna’s Archive doesn’t come with a paywall or shopping cart.
“We strongly believe in the free flow of information, and preservation of knowledge and culture. With this search engine, we build on the shoulders of giants,” Anna writes.
The site isn’t completely finished yet and may still have quite a few bugs. But given the situation with Z-Library, the team wanted to get the search engine up and running as soon as possible.
The Z-Library links rely on the Tor version of the site, which remains online. However, the goal is to ultimately make all content available through IPFS as well. This would make it pretty much impossible to take down, similar to the Library Genesis forks, which also use IPFS.
Anna informed TorrentFreak that the Z-Library domain seizures came as a shock but not a deterrent. They only increased the team’s motivation to go full steam ahead, while being mindful of their privacy.
“For us, it underscores the importance of being careful, and the importance of this data being widely available and redundant across the globe,” Anna said.
The privacy angle is a topic that has become very relevant this week. Details of the criminal indictment of Z-Library’s alleged operators show that law enforcement has a lot of tools to track people down, even those who prefer to remain anonymous.
Anna and the team know the stakes. In a recent post, she wrote that Sci-Hub founder Alexandra Elbakyian, whose identity became public, risks being arrested if she travels to a Western country. This is why Anna’s team does all it can to remain anonymous.
“We are at the other end of the spectrum; being very careful not to leave any trace, and having strong operational security,” she wrote.
We asked a series of questions to find out more about the motivations behind the project but Anna says they’re not ready to do interviews just yet. For now, the team will focus on its archival ‘duties’.
“We are planning to write an in-depth essay about our motivations at some point,” she said, noting that in-depth interviews may be an option later.
From: TF, for the latest news on copyright battles, piracy and more.
Chinese phone maker Xiaomi has released a couple of foldable phones in the past few years. But the company has been showing off prototypes for even longer… including for some models that never make it to market. A few years ago Xiaomi filed for …
Chinese phone maker Xiaomi has released a couple of foldable phones in the past few years. But the company has been showing off prototypes for even longer… including for some models that never make it to market. A few years ago Xiaomi filed for a patent on a foldable that bends outward rather than inward, […]
The post Lilbits: Prototype of an unreleased Xiaomi foldable spotted appeared first on Liliputing.
ID.me said $414 billion was lost—10 times higher than feds’ estimate.
Enlarge / House Oversight Committee Chair Rep. Carolyn Maloney (D-NY) is seen during a hearing in March 2022 in Washington, DC. (credit: Alex Wong / Staff | Getty Images News)
Online verification through ID.me was supposed to help the Internal Revenue Service (IRS) prevent widespread fraud while rushing unemployment benefits to those who needed it most during the pandemic.
Instead, wait times to get verified for some stretched up to 10 hours, and in one case in California, actually took three days. A six-month investigation by the House Oversight Committee found evidence that ID.me “downplayed excessive wait times,” hiding the problem from federal and state agencies by only providing data related to successful connections, Politico reported. This likely kept aid out of reach for those most in need, whose technology access was not reliable enough to stay connected through substantially delayed wait times.
In a statement, chairwoman of the House Committee on Oversight and Reform, Representative Carolyn Maloney (D-NY), described the evidence revealed by the investigation as “appalling.” The Select Subcommittee on the Coronavirus Crisis chairman, Rep. James Clyburn (D-SC), expressed disappointment that ID.me did not deliver as a partner in US efforts to quickly disburse aid.
You must be logged in to post a comment.