Microsoft’s new Security Copilot will help network admins respond to threats in minutes, not day

Posted on by

[…]

with Microsoft’s unveiling of the new Security Copilot AI at its inaugural Microsoft Secure event. The automated enterprise-grade security system is powered by OpenAI’s GPT-4, runs on the Azure infrastructure and promises admins the ability “to move at the speed and scale of AI.”

Security Copilot is similar to the large language model (LLM) that drives the Bing Copilot feature, but with a training geared heavily towards network security rather than general conversational knowledge and web search optimization. […]

“Just since the pandemic, we’ve seen an incredible proliferation [in corporate hacking incidents],”Jakkal told Bloomberg. For example, “it takes one hour and 12 minutes on average for an attacker to get full access to your inbox once a user has clicked on a phishing link. It used to be months or weeks for someone to get access.”

[…]

Jakkal anticipates these new capabilities enabling Copilot-assisted admins to respond within minutes to emerging security threats, rather than days or weeks after the exploit is discovered. Being a brand new, untested AI system, Security Copilot is not meant to operate fully autonomously, a human admin needs to remain in the loop. “This is going to be a learning system,” she said. “It’s also a paradigm shift: Now humans become the verifiers, and AI is giving us the data.”

To more fully protect the sensitive trade secrets and internal business documents Security Copilot is designed to protect, Microsoft has also committed to never use its customers data to train future Copilot iterations. Users will also be able to dictate their privacy settings and decide how much of their data (or the insights gleaned from it) will be shared. The company has not revealed if, or when, such security features will become available for individual users as well.

Source: Microsoft’s new Security Copilot will help network admins respond to threats in minutes, not days | Engadget

Inaudible ultrasound attack can stealthily control your phone, smart speaker

Posted on by

American university researchers have developed a novel attack called “Near-Ultrasound Inaudible Trojan” (NUIT) that can launch silent attacks against devices powered by voice assistants, like smartphones, smart speakers, and other IoTs.

The team of researchers consists of professor Guenevere Chen of the University of Texas in San Antonio (UTSA), her doctoral student Qi Xia, and professor Shouhuai Xu of the University of Colorado (UCCS).

The team demonstrated NUIT attacks against modern voice assistants found inside millions of devices, including Apple’s Siri, Google’s Assistant, Microsoft’s Cortana, and Amazon’s Alexa, showing the ability to send malicious commands to those devices.

Inaudible attacks

The main principle that makes NUIT effective and dangerous is that microphones in smart devices can respond to near-ultrasound waves that the human ear cannot, thus performing the attack with minimal risk of exposure while still using conventional speaker technology.

In a post on USTA’s site, Chen explained that NUIT could be incorporated into websites that play media or YouTube videos, so tricking targets into visiting these sites or playing malicious media on trustworthy sites is a relatively simple case of social engineering.

The researchers say the NUIT attacks can be conducted using two different methods.

The first method, NUIT-1, is when a device is both the source and target of the attack. For example, an attack can be launched on a smartphone by playing an audio file that causes the device to perform an action, such as opening a garage door or sending a text message.

The other method, NUIT-2, is when the attack is launched by a device with a speaker to another device with a microphone, such as a website to a smart speaker.

Source: Inaudible ultrasound attack can stealthily control your phone, smart speaker

This is like smart units like Amazon Echo / Alexa being controlled by TV commercials

Meatball from long-extinct mammoth created by food firm

Posted on by

A mammoth meatball has been created by a cultivated meat company, resurrecting the flesh of the long-extinct animals.

The project aims to demonstrate the potential of meat grown from cells, without the slaughter of animals, and to highlight the link between large-scale livestock production and the destruction of wildlife and the climate crisis.

The mammoth meatball was produced by Vow Food, an Australian company, which is taking a different approach to cultured meat.

There are scores of companies working on replacements for conventional meat, such as chicken, pork and beef. But Vow Foods is aiming to mix and match cells from unconventional species to create new kinds of meat.

The company has already investigated the potential of more than 50 species, including alpaca, buffalo, crocodile, kangaroo, peacocks and different types of fish.

The first cultivated meat to be sold to diners will be Japanese quail, which the company expects will be in restaurants in Singapore this year.

“We have a behaviour change problem when it comes to meat consumption,” said George Peppou, CEO of Vow Food.

“The goal is to transition a few billion meat eaters away from eating [conventional] animal protein to eating things that can be produced in electrified systems.

“And we believe the best way to do that is to invent meat. We look for cells that are easy to grow, really tasty and nutritious, and then mix and match those cells to create really tasty meat.”

Tim Noakesmith, who cofounded Vow Food with Peppou, said: “We chose the woolly mammoth because it’s a symbol of diversity loss and a symbol of climate change.” The creature is thought to have been driven to extinction by hunting by humans and the warming of the world after the last ice age.

[…]

Cultivated meat – chicken from Good Meat – is currently only sold to consumers in Singapore, but two companies have now passed an approval process in the US.

In 2018, another company used DNA from an extinct animal to create gummy bears made from gelatine from a mastodon, another elephant-like animal.

Vow Food worked with Prof Ernst Wolvetang, at the Australian Institute for Bioengineering at the University of Queensland, to create the mammoth muscle protein. His team took the DNA sequence for mammoth myoglobin, a key muscle protein in giving meat its flavour, and filled in the few gaps using elephant DNA.

This sequence was placed in myoblast stem cells from a sheep, which replicated to grow to the 20bn cells subsequently used by the company to grow the mammoth meat.

“It was ridiculously easy and fast,” said Wolvetang. “We did this in a couple of weeks.” Initially, the idea was to produce dodo meat, he said, but the DNA sequences needed do not exist.

[…]

Seren Kell, at the Good Food Institute Europe, said: “I hope this fascinating project will open up new conversations about cultivated meat’s extraordinary potential to produce more sustainable food.

“However, as the most common sources of meat are farm animals such as cattle, pigs, and poultry, most of the sustainable protein sector is focused on realistically replicating meat from these species.

“By cultivating beef, pork, chicken and seafood we can have the most impact in terms of reducing emissions from conventional animal agriculture.”

[…]

Source: Meatball from long-extinct mammoth created by food firm | Meat industry | The Guardian

Italian art experts (and rest of world) astonished by David statue uproar in Florida

Posted on by

The Florence museum that houses Michelangelo’s statue of David has invited teachers and students from a Florida school to visit, after an uproar over an art lesson.

The school’s principal quit after a complaint about a sixth-grade art class that included an image of the statue.

A parent had complained the image was pornographic.

Cecilie Hollberg, director of Galleria dell’Accademia, has now issued the invitation to the class.

She said the principal should be “rewarded, not punished”.

“Talking about the Renaissance without showing the David, an undisputed icon of art and culture and of that historical period, would make no sense,” Ms Hollberg said.

The controversy began when the board of Tallahassee Classical School – a charter school in Florida’s state capital – pressured principal Hope Carrasquilla to resign after three parents complained about a lesson that included a photo of the 17ft nude marble statue.

The statue, one of the most famous in Western history, depicts the biblical David going to fight Goliath armed only with a sling and his faith in God.

[…]

The incident has left Florentines and experts on Renaissance art bewildered.

The David is considered a masterpiece of the Italian Renaissance and a symbol of humanist values. It has been displayed in the Galleria dell’Accademia since 1873.

Ms Hollberg said she was “astonished”, stating that to think that the David statue could be considered pornographic means not only failing to understand the Bible, but Western culture itself.

“I cannot believe that actually happened, at first I thought it was fake news, so improbable and absurd was it,” she said.

“A distinction must be made between nudity and pornography. There is nothing pornographic or aggressive about the David, he is a young boy, a shepherd, who even according to the Bible did not have ostentatious clothes but wanted to defend his people with what he had.”

[…]

According to Florentine art historian and dean of the University for Foreigners in Siena, Tomaso Montanari, such an attitude is “disconcerting”.

“First comes the dismay at the absence of educational freedom, as it should not be restricted or manipulated by families,” Mr Montanari said.

“On the other hand, from a cultural perspective, the Western world has a tendency to associate fundamentalism and censorship with other societies, believing it possesses the capability to spread democratic ideals worldwide.

“But this cultural backsliding clearly highlights the presence of fundamentalist views within the West as well.”

[…]

Back in Florence, Ms Hollberg remarked: “From majestic statues to charming fountains and paintings, Italy is overflowing with works of art, not just in its museums, but in all its cities, squares and streets, with some featuring naked figures.

“Does that make it pornography? Should entire cities be shut down because of the artistic depictions of the human form?”

Source: Italian art experts astonished by David statue uproar in Florida – BBC News

ChatGPT Retrieval plugin allows you to embed and store memory of your chats and other knowledge bases

Posted on by

This is a plugin for ChatGPT that enables semantic search and retrieval of personal or organizational documents. It allows users to obtain the most relevant document snippets from their data sources, such as files, notes, or emails, by asking questions or expressing needs in natural language. Enterprises can make their internal documents available to their employees through ChatGPT using this plugin.

[…]

Users can refine their search results by using metadata filters by source, date, author, or other criteria. The plugin can be hosted on any cloud platform that supports Docker containers, such as Fly.io, Heroku or Azure Container Apps. To keep the vector database updated with the latest documents, the plugin can process and store documents from various data sources continuously

[…]

A notable feature of the Retrieval Plugin is its capacity to provide ChatGPT with memory. By utilizing the plugin’s upsert endpoint, ChatGPT can save snippets from the conversation to the vector database for later reference (only when prompted to do so by the user). This functionality contributes to a more context-aware chat experience by allowing ChatGPT to remember and retrieve information from previous conversations. Learn how to configure the Retrieval Plugin with memory here.

[…]

Github openai/chatgpt-retreival-plugin

‘A Blow for Libraries’: Internet Archive Loses Copyright Infringement Lawsuit by money grubbing publishers

Posted on by

A judge ruled against Internet Archive, a free online digital library, on Friday in a lawsuit filed by four top publishers who claimed the company was in violation of copyright laws. The publishers, Hachette Book Group, HarperCollins, John Wiley & Sons, and Penguin Random House filed the lawsuit against Internet Archive in 2020, claiming the company had illegally scanned and uploaded 127 of their books for readers to download for free, detracting from their sales and the authors’ royalties.

U.S. District Court Judge John G. Koeltl ruled in favor of the publishing houses in saying that Internet Archive was making “derivative” works by transforming printed books into e-books and distributing them.

[…]

Koeltl’s decision was in part based on the law that libraries are required to pay publishers for continued use of their digital book copies and are only permitted to lend these digital copies a specified number of times, called controlled digital lending, as agreed by the publisher [not the writer!] before paying to renew its license.

[…]

However, according to the court ruling, Hachette and Penguin provide one or two-year terms to libraries, in which the eBook can be rented an unlimited number of times before the library has to purchase a new license. HarperCollins allows the library to circulate a digital copy 26 times before the license has to be renewed, while Wiley has continued to experiment with several subscription models.

[…]

The judge ruled that because Internet Archive was purchasing the book only once before scanning it and lending each digital copy an unlimited number of times, it is an infringement of copyright and “concerns the way libraries lend eBooks.

[…]

Source: ‘A Blow for Libraries’: Internet Archive Loses Copyright Infringement Lawsuit

The decision was “a blow to all libraries and the communities we serve,” argued Chris Freeland, the director of Open Libraries at the Internet Archive. In a blog post he argued the decision “impacts libraries across the U.S. who rely on controlled digital lending to connect their patrons with books online. It hurts authors by saying that unfair licensing models are the only way their books can be read online. And it holds back access to information in the digital age, harming all readers, everywhere.
The Verge adds that the judge rejected “fair use” arguments which had previously protected a 2014 digital book preservation project by Google Books and HathiTrust:

Koetl wrote that any “alleged benefits” from the Internet Archive’s library “cannot outweigh the market harm to the publishers,” declaring that “there is nothing transformative about [Internet Archive’s] copying and unauthorized lending,” and that copying these books doesn’t provide “criticism, commentary, or information about them.” He notes that the Google Books use was found “transformative” because it created a searchable database instead of simply publishing copies of books on the internet.

Their lending model works like this. They purchase a paper copy of the book, scan it to digital format, and then lend out the digital copy to one person at a time. Their argument is that this is no different than lending out the paper copy that they legally own to one person at a time. It is not as cut and dry as you make it out to be.

Source: Internet Archive Loses in Court. Judge Rules They Can’t Scan and Lend eBooks

Last Monday was the day of the oral arguments in the Big Publishers’ lawsuit against libraries in the form of the Internet Archive. As we noted mid-week, publishers won’t quit until libraries are dead. And they got one step closer to that goal on Friday, when Judge John Koetl wasted no time in rejecting every single one of the Internet Archive’s arguments.

The fact that the ruling came out on the Friday after the Monday oral arguments suggests pretty strongly that Judge Koetl had his mind made up pretty quickly and was ready to kill a library with little delay. Of course, as we noted just last Wednesday, whoever lost at this stage was going to appeal, and the really important stuff was absolutely going to happen at the 2nd Circuit appeals court. It’s just that now the Internet Archives, and a bunch of important copyright concepts, are already starting to be knocked down a few levels.

I’ve heard from multiple people claiming that of course the Internet Archive was going to lose, because it was scanning books (!!) and lending them out and how could that be legal? But, the answer, as we explained multiple times, is that every piece of this copyright puzzle had already been deemed legal.

And the Internet Archive didn’t just jump into this without any thought. Two of the most well known legal scholars regarding copyright and libraries, David Hansen and Kyle Courtney, had written a white paper detailing exactly how and why the approach the Internet Archive took with Controlled Digital Lending easily fit within the existing contours and precedents of copyright law.

But, as we and others have discussed for ages, in the copyright world, there’s a long history of courts ignoring what the law actually says and just coming up with some way to say something is infringement if it feels wrong to them. And that’s what happened here.

A key part of the ruling, as in a large percentage of cases that are about fair use, is looking at whether or not the use of the copy is “transformative.” Judge Koetl is 100% positive it is not transformative.

There is nothing transformative about IA’s copying and unauthorized lending of the Works in Suit.7 IA does not reproduce the Works in Suit to provide criticism, commentary, or information about them. See 17 U.S.C. § 107. IA’s ebooks do not “add[] something new, with a further purpose or different character, altering the [originals] with new expression, meaning or message.” Campbell, 510 U.S. at 579. IA simply scans the Works in Suit to become ebooks and lends them to users of its Website for free. But a copyright holder holds the “exclusive[] right” to prepare, display, and distribute “derivative works based upon the copyrighted work.”

But… there’s a lot more to “transformative” use than simply adding something new or altering the meaning. In many cases, fair use is found in cases where you’re copying the exact same content, but for a different purpose, and the Internet Archive’s usage here seems pretty clearly transformative in that it’s changing the way the book can be consumed to make it easier for libraries to lend it out and patrons to read it. That is, the “transformation” is in the way the book can be lent, not the content of the book.

I know many people find this strange (and the judge did here as well) saying things like “but it’s the whole work.” Or “the use is the same because it’s still just reading the book.” But the Supreme Court already said, quite clearly, that such situations can be fair use, such as in the Sony v. Universal case that decided VCRs were legal, and that time shifting TV shows was clear fair use. In that ruling, they even cite Congress noting that “making a copy of a copyright work for… convenience” can be considered fair use.

Unfortunately, Judge Koetl effectively chops away a huge part of the Sony ruling in insisting that this is somehow different.

But Sony is plainly inapposite. IA is not comparable to the parties in Sony — either to Sony, the alleged contributory copyright infringer, or to the home viewers who used the Betamax machine for the noncommercial, nonprofit activity of watching television programs at home. Unlike Sony, which only sold the machines, IA scans a massive number of copies of books and makes them available to patrons rather than purchasing ebook licenses from the Publishers. IA is also unlike the home viewers in Sony, who engaged in the “noncommercial, nonprofit activity” of viewing at a more convenient time television programs that they had the right to view for free at the time they were originally broadcast. 464 U.S. at 449. The home viewers were not accused of making their television programs available to the general public. Although IA has the right to lend print books it lawfully acquired, it does not have the right to scan those books and lend the digital copies en masse.

But note what the Judge did here. Rather than rely on the text of what the Supreme Court actually said in Sony, he insists that he won’t apply the rules of Sony because the parties are different. But if the basic concepts and actions are covered by the Sony ruling, it seems silly to ignore them here as the judge did.

And the differences highlighted by the court here have no bearing on the actual issues and the specifics of fair use and the law. I mean, first of all, the fact that Koetl claims that the Internet Archive is not engaged in “noncommercial, nonprofit activity” is just weird. The Internet Archive is absolutely engaged in noncommerical, nonprofit activity.

The other distinctions are meaningless as well. No, IA is not building devices for people to buy, but in many ways IA’s position here should be seen as stronger than Sony’s because Sony actually was a commercial operation, and IA is literally acting as a library, increasing the convenience for its patrons, and doing so in a manner that is identical to lending out physical books. Sony created a machine, Betamax, that copied TV shows and allowed those who bought those machines to watch the show at a more convenient time. IA created a machine that copies books, and allows library patrons to access those books in a more convenient way.

Also, the Betamax (and VCR) were just as “available to the general public” as the Internet Archive is. The idea that they are substantially different is just… weird. And strikes me as pretty clearly wrong.

There’s another precedential oddity in the ruling. It relies pretty heavily on the somewhat terrible fair use ruling in the 2nd Circuit in the Warhol Foundation v. Goldsmith case. That case was so terrible that we (at the Copia Institute) weighed in with the Supreme Court to let them know how problematic it was, and the Supreme Court is still sitting on a decision in that case.

Which means the Supreme Court is soon to rule on it, and that could very much change or obliterate the case that Judge Koetl leans on heavily for his ruling.

Here, Judge Koetl repeatedly goes back to the Warhol well to make various arguments, especially around the question of the fourth fair use factor: the effect on the market. To me, this clearly weighs towards fair use, because it’s no different than a regular library. Libraries are allowed to buy (or receive donated) books and lend them out. That’s all the Open Library does here. So to argue there’s a negative impact on the market, the publishers rely on the fact that they’ve been able to twist and bend copyright law so much that they’ve created a new, extortionate market in ebook “licenses,” and then play all sorts of games to force people to buy the books rather than check them out of the library.

Judge Koetl seems particularly worried about how much damage this could do this artificially inflated market:

It is equally clear that if IA’s conduct “becomes widespread, it will adversely affect the potential market for the” Works in Suit. Andy Warhol Found., 11 F.4th at 48. IA could expand the Open Libraries project far beyond the current contributing partners, allowing new partners to contribute many more concurrent copies of the Works in Suit to increase the loan count. New organizations like IA also could emerge to perform similar functions, further diverting potential readers and libraries from accessing authorized library ebooks from the Publishers. This plainly risks expanded future displacement of the Publishers’ potential revenues.

But go back and read that paragraph again, and replace the key words to read that if libraries become widespread, it will adversely affect the potential market for buying books in bookstores… because libraries would be “diverting potential readers” from purchasing physical books, which “plainly risks expanded future displacement of the Publishers’ potential revenues.”

Again, the argument here is effectively that libraries themselves shouldn’t be allowed. And that seems like a problem?

Koetl also falls into the ridiculous trap of saying that “you can’t compete with free” and that libraries will favor CDL-scanned books over licensed ones:

An accused infringer usurps an existing market “where the infringer’s target audience and the nature of the infringing content is the same as the original.” Cariou, 714 F.3d at 709; see also Andy Warhol Found., 11 F.4th at 50. That is the case here. For libraries that are entitled to partner with IA because they own print copies of books in IA’s collection, it is patently more desirable to offer IA’s bootleg ebooks than to pay for authorized ebook licenses. To state the obvious, “[i]t is difficult to compete with a product offered for free.” Sony BMG Music Ent. v. Tenenbaum, 672 F. Supp. 2d 217, 231 (D. Mass. 2009).

Except that’s literally wrong. The licensed ebooks have many features that the scanned ones don’t. And many people (myself included!) prefer to check out licensed ebooks from our local libraries rather than the CDL ones, because they’re more readable. My own library offers the ability to check out books from either one, and defaults to recommending the licensed ebooks, because they’re a better customer experience, which is how tons of products “compete with free” all the time.

I mean, not to be simplistic here, but the bottled water business in the US is an over $90 billion market for something most people can get for free (or effectively free) from the tap. That’s three times the size of the book publishing market. So, uh, maybe don’t say “it’s difficult to compete with free.” Other industries do it just fine. The publishers are just being lazy.

Besides, based on this interpretation of Warhol, basically anyone can destroy fair use by simply making up some new, crazy, ridiculously priced, highly restrictive license that covers the same space as the fair use alternative, and claim that the alternative destroys the “market” for this ridiculous license. That can’t be how fair use works.

Anyway, one hopes first that the Supreme Court rejects the terrible 2nd Circuit ruling in the Warhol Foundation case, and that this in turn forces Judge Koetl to reconsider his argument. But given the pretzel he twisted himself into to ignore the Betamax case, it seems likely he’d still find against libraries like the Internet Archive.

Given that, it’s going to be important that the 2nd Circuit get this one right. As the Internet Archive’s Brewster Kahle said in a statement on the ruling:

“Libraries are more than the customer service departments for corporate database products. For democracy to thrive at global scale, libraries must be able to sustain their historic role in society—owning, preserving, and lending books.

This ruling is a blow for libraries, readers, and authors and we plan to appeal it.”

What happens next is going to be critical to the future of copyright online. Already people have pointed out how some of the verbiage in this ruling could have wide reaching impact on questions about copyright in generative AI products or many other kinds of fair use cases.

One hopes that the panel on the 2nd Circuit doesn’t breezily dismiss these issues like Judge Koetl did.

Source: Publishers Get One Step Closer To Killing Libraries

This money grab by publishers is disgusting, for more information and articles I have referenced them here

Apple acquired a startup using AI to compress videos

Posted on by

Apple has quietly acquired a Mountain View-based startup, WaveOne, that was developing AI algorithms for compressing video.

Apple wouldn’t confirm the sale when asked for comment. But WaveOne’s website was shut down around January, and several former employees, including one of WaveOne’s co-founders, now work within Apple’s various machine learning groups.

WaveOne’s former head of sales and business development, Bob Stankosh, announced the sale in a LinkedIn post published a month ago.

“After almost two years at WaveOne, last week we finalized the sale of the company to Apple,” Stankosh wrote. “We started our journey at WaveOne, realizing that machine learning and deep learning video technology could potentially change the world. Apple saw this potential and took the opportunity to add it to their technology portfolio.”

[…]

WaveOne’s main innovation was a “content-aware” video compression and decompression algorithm that could run on the AI accelerators built into many phones and an increasing number of PCs. Leveraging AI-powered scene and object detection, the startup’s technology could essentially “understand” a video frame, allowing it to, for example, prioritize faces at the expense of other elements within a scene to save bandwidth.

WaveOne also claimed that its video compression tech was robust to sudden disruptions in connectivity. That is to say, it could make a “best guess” based on whatever bits it had available, so when bandwidth was suddenly restricted, the video wouldn’t freeze; it’d just show less detail for the duration.

WaveOne claimed its approach, which was hardware-agnostic, could reduce the size of video files by as much as half, with better gains in more complex scenes.

[…]

Even minor improvements in video compression could save on bandwidth costs, or enable services like Apple TV+ to deliver higher resolutions and framerates depending on the type of content being streamed.

YouTube’s already doing this. Last year, Alphabet’s DeepMind adapted a machine learning algorithm originally developed to play board games to the problem of compressing YouTube videos, leading to a 4% reduction in the amount of data the video-sharing service needs to stream to users.

[…]

Source: Apple acquired a startup using AI to compress videos | TechCrunch

Nike Blocks F1 Champ Max Verstappen’s ‘Max 1’ Clothing Brand because they can own words now

Posted on by

[…]

Nike’s argument is that Max 1 is too similar to its longtime “Air Max” shoe line, including other “Max Force 1” products and other variations that include similar keywords. Verstappen had named his line of products after himself and his current racing number but encountered legal trouble soon after launch.

The Benelux Office for Intellectual Property—essentially The Netherlands’ trademark office—issued a report that Verstappen’s Max 1 brand carried a “likelihood of confusion” and posed a risk of consumers believing Max 1 products were associated with Nike.

Nike went as far as claiming that some designs in the Max 1 catalog were too similar to the apparel giant’s, while also alleging that the word MAX was prominently used and likened to Nike apparel. For these reasons, Verstappen was reportedly fined approximately $1,100 according to Express.

[…]

Source: Nike Blocks F1 Champ Max Verstappen’s ‘Max 1’ Clothing Brand

1. What about Pepsi Max

2. What about the name Max being much much older than Nike (so prior use)

3. What about people going around using the word max as in eg ‘that’s the max speed’ or ‘that’s the max that will go in’?

3. What the actual fuck, copyright.

U.S. Political Party Preferences Shifted Greatly During 2021

Posted on by

[…]The general stability for the full-year average obscures a dramatic shift over the course of 2021, from a nine-percentage-point Democratic advantage in the first quarter to a rare five-point Republican edge in the fourth quarter.

Line graph. Quarterly averages of U.S. party identification and leaning in 2021. In the first quarter of 2021, 49% of U.S. adults identified as Democrats or leaned Democratic, while 40% identified as Republicans or leaned Republican. In the second quarter, 49% were Democrats or Democratic leaners, and 43% were Republicans and Republican leaners. In the third quarter, 45% were Democrats and Democratic leaners, and were 44% Republicans and Republican leaners. In the fourth quarter, 42% were Democrats and Democratic leaners, and 47% were Republicans and Republican leaners.

These results are based on aggregated data from all U.S. Gallup telephone surveys in 2021, which included interviews with more than 12,000 randomly sampled U.S. adults.

[…]

Shifting party preferences in 2021 are likely tied to changes in popularity of the two men who served as president during the year.

[…]

With Trump’s approval rating at a low point and Biden relatively popular in the first quarter, 49% of Americans identified as Democrats or leaned Democratic, compared with 40% who were Republicans or Republican leaners.

In the second quarter, Democratic affiliation stayed high, while Republican affiliation began to recover, increasing to 43%.

The third quarter saw a decline in Democratic identification and leaning, from 49% to 45%, as Biden’s ratings began to falter, while there was no meaningful change in Republican affiliation.

In the fourth quarter, party support flipped as Republicans made gains, from 44% to 47%, and Democratic affiliation fell from 45% to 42%. These fourth-quarter shifts coincided with strong GOP performances in 2021 elections, including a Republican victory in the Virginia gubernatorial election and a near-upset of the Democratic incumbent governor in New Jersey. Biden won both states by double digits in the 2020 election.

[…]

The shifts in party affiliation in each quarter of 2021 were apparent in both the percentage identifying with each party and the percentage of independents leaning to each party, but with more changes among leaners than identifiers.

[…]

Republican identification increased by three points from the beginning to the end of 2021, while Republican leaners increased by four points.

Changes in Party Identification and Leaning, by Quarter, 2021
In politics, as of today, do you consider yourself — [a Republican, a Democrat] — or an independent? // As of today, do you lean more to the — [Democratic Party or the Republican Party]?
2021-I 2021-II 2021-III 2021-IV
% % % %
Democrat 30 31 28 28
Democratic-leaning independent 19 18 17 14
Non-leaning independent 10 5 8 9
Republican-leaning independent 15 17 16 19
Republican 25 26 28 28
Percentage with no opinion not shown
Gallup

[…]

Independents Are Still the Largest Political Group in the U.S.

Regardless of which party has an advantage in party affiliation, over the past three decades, presidential elections have generally been competitive, and party control of the U.S. House of Representatives and Senate has changed hands numerous times. This is partly because neither party can claim a very high share of core supporters — those who identify with the party — as the largest proportion of Americans identify initially as political independents.

Overall in 2021, an average of 29% of Americans identified as Democrats, 27% as Republicans and 42% as independents. Roughly equal proportions of independents leaned to the Democratic Party (17%) and to the Republican Party (16%).

The percentage of independent identifiers is up from 39% in 2020, but similar to the 41% measured in 2019. Gallup has often seen a decrease in independents in a presidential election year and an increase in the year after.

[…]

[…]

Source: U.S. Political Party Preferences Shifted Greatly During 2021

Small subgroups of the population seem much larger to many Americans and they think Large groups are smaller than they really are

Posted on by

When it comes to estimating the size of demographic groups, Americans rarely get it right. In two recent YouGov polls, we asked respondents to guess the percentage (ranging from 0% to 100%) of American adults who are members of 43 different groups,[…] When people’s average perceptions of group sizes are compared to actual population estimates, an intriguing pattern emerges: Americans tend to vastly overestimate the size of minority groups.

[…]

A parallel pattern emerges when we look at estimates of majority groups: People tend to underestimate rather than overestimate their size relative to their actual share of the adult population.

[…]

The most accurate estimates involved groups whose real proportion fell right around 50%

[…]

Although there is some question-by-question variability, the results from our survey show that inaccurate perceptions of group size are not limited to the types of socially charged group divisions typically explored in similar studies: race, religion, sexuality, education, and income. Americans are equally likely to misestimate the size of less widely discussed groups, such as adults who are left-handed.

[…]

Similar misperceptions are found regarding the proportion of American adults who own a pet, have read a book in the past year, or reside in various cities or states. This suggests that errors in judgment are not due to the specific context surrounding a certain group.

[…]

Source: ​​From millionaires to Muslims, small subgroups of the population seem much larger to many Americans | YouGov

New Zealand’s Dawn Aerospace Mk-II Aurora Space drone Approved for Suborbital Test Flights

Posted on by

Dawn Aerospace CEO Stefan Powell announced today that the company’s Mk-II Aurora spaceplane has received approval from the Civil Aviation Authority of New Zealand for rocket-powered flight. The company is now ready to test the vehicle’s rocket engines with flights beginning next month.

The Mk-II Aurora is a remotely piloted spaceplane that could eventually take two trips into space every day. During testing, the Mk-II will host research projects and collect scientific data with its onboard 3U payload capacity (i.e. 30 cubic centimeters in size) while serving as a proof of concept for a later model—the Mk-III—that could deliver 550-pound (250-kilogram) satellites into orbit with the help of a second stage rocket (the concept is somewhat similar to how Virgin Orbit uses piloted aircraft to deploy its LauncherOne rocket).

With approval from New Zealand’s Civil Aviation Authority, Dawn Aerospace is readying to begin rocket-powered flight tests of the Mk-II Aurora beginning in just a few weeks. The company has already fired the Aurora’s rocket engine—112 times according to Powell in a press release—but never has it conducted a flight test with the engine, opting to flight test 48 times with jet engines instead.

[…]

tests of the Mk-II Aurora will follow a “build-up approach,” as previous testing has, in which the plane will reach “modest” altitudes and speeds to demonstrate that rocket-powered flight is just as viable as commercial aircraft.

The Mk-II Aurora.
The Mk-II Aurora.
Photo: Dawn Aerospace

The altitude the Aurora will reach in these upcoming tests isn’t necessarily modest, as company plans to reach the internationally recognized boundary of space, better known as the Karman line. As SatNews reported last September, Dawn Aerospace is gearing up for Phase 2 of testing the Aurora, which according to its company’s website, will see the company push Aurora to higher and higher altitudes over a series of tests until it crosses the Karman line, which is about 62 miles (100 kilometers) above the Earth’s surface.

Source: New Zealand’s Spaceplane Approved for Suborbital Test Flights

GitHub.com rotates its exposed private SSH key

Posted on by

GitHub has rotated its private SSH key for GitHub.com after the secret was was accidentally published in a public GitHub repository.

The software development and version control service says, the private RSA key was only “briefly” exposed, but that it took action out of “an abundance of caution.”

Unclear window of exposure

In a succinct blog post published today, GitHub acknowledged discovering this week that the RSA SSH private key for GitHub.com had been ephemerally exposed in a public GitHub repository.

“We immediately acted to contain the exposure and began investigating to understand the root cause and impact,” writes Mike Hanley, GitHub’s Chief Security Officer and SVP of Engineering.

“We have now completed the key replacement, and users will see the change propagate over the next thirty minutes. Some users may have noticed that the new key was briefly present beginning around 02:30 UTC during preparations for this change.”

The timing of the discovery is interesting—just weeks after GitHub rolled out secrets scanning for all public repos.

GitHub.com’s latest public key fingerprints are shown below. These can be used to validate that your SSH connection to GitHub’s servers is indeed secure.

As some may notice, only GitHub.com’s RSA SSH key has been impacted and replaced. No change is required for ECDSA or Ed25519 users.

SHA256:uNiVztksCsDhcc0u9e8BujQXVUpKZIDTMczCvj3tD2s (RSA)
SHA256:br9IjFspm1vxR3iA35FWE+4VTyz1hYVLIE2t1/CeyWQ (DSA – deprecated)
SHA256:p2QAMXNIC1TJYWeIOttrVc98/R1BUFWu3/LiyKgUfQM (ECDSA)
SHA256:+DiY3wvvV6TuJJhbpZisF/zLDA0zPMSvHdkr4UvCOqU (Ed25519)

“Please note that this issue was not the result of a compromise of any GitHub systems or customer information,” says GitHub.

“Instead, the exposure was the result of what we believe to be an inadvertent publishing of private information.”

The blog post, however, does not answer when exactly was the key exposed, and for how long, making the timeline of exposure a bit murky. Such timestamps can typically be ascertained from security logs—should these be available, and Git commit history.

[…]

Source: GitHub.com rotates its exposed private SSH key

Nordic Air Defense Pact Combines Forces Of Hundreds Of Fighter Aircraft

Posted on by

To better cope with threats emanating from Russia, the countries of Denmark, Finland, Norway and Sweden have created a unified Nordic air defense alliance, pooling the resources of their air forces. They have upwards of 300 fighter jets between them as well as training, transport and surveillance fixed-wing aircraft and helicopters.

Those four nations on Friday announced they signed the first Nordic Air Commanders’ Intent last week during a meeting at Ramstein Air Base in Germany.

“The declaration of intent strengthens Nordic cooperation and paves the way for further strengthening of the Nordic air forces,” the four nations said Friday in a joint statement. “The ultimate goal is to be able to operate seamlessly together as one force by developing a Nordic concept for joint air operations based on already-known NATO methodology.”

To achieve that goal, this intent directs the development of a “Nordic Warfighting Concept for Joint Air Operations,” pursuing four lines of effort:

  • integrated command and control, operational planning and execution
  • flexible and resilient deployment of our air forces
  • joint airspace surveillance
  • joint education, training and exercises.

The publicly released plan does not provide specific timelines for achieving any of the goals. However, a separate jointly released document gives an overview.

Finnish Air Force F-18 Hornet. (Finnish Air Force photo)

“In the medium term, efforts shall revolve around preparing for, conducting, and assessing Nordic Response 24 from an air perspective, putting emphasis on the Nordic digital and semi-distributed [Air Operations Center] AOC development steps,” according to that document. “On the horizon, long-term permanent solutions to fulfill this intent’s aim shall be determined and established.”

While none of the documents mention Russia, the move to integrate the air forces was triggered by Moscow’s full-on invasion of Ukraine, the commander of the Danish Air Force, Major General Jan Dam, told Reuters.

“Our combined fleet can be compared to a large European country,” Dam said.

Norway has at least 52 F-35 Lightning II Joint Strike Fighters, according to Janes. The Norwegian Air Force says it is phased out its F-16 fleet.

Norway flies F-35 Lightning II Joint Strike Fighters. (Norwegian Air Force photo)

Finland has 62 F/A-18C/D multirole fighter jets and 64 F-35s on order, according to Reuters.

Finland has more than 60 Hornets. (Finnish Air Force photo)

Finnish Defense Minister Antti Kaikkonen on Thursday expressed his opposition to a request by Ukraine for a portion of his country’s Hornet fleet.

“My view as Finland’s defense minister is that we need these Hornets to secure our own country,” Kaikkonen told a news conference in Helsinki, as reported by Reuters. “I view negatively the idea that they would be donated during the next few years. And if we look even further, my understanding is that they begin to be worn out and will have little use value left, he added.”

Denmark has 58 F-16s and 27 F-35s on order, according to Reuters.

Danish F-16 taxiing ready for a training mission alongside Allies in the Baltic Sea region, helping improve tactics and readiness. (Danish Air Force photo).

Sweden has around 70 JAS-39C/D Gripen jets and will be converting over to the enhanced Gipen-E in the coming years.

Swedish JAS-38 Gripen jets. (Swedish Air Force photo)

How soon this gets off the ground and exactly how it will works remains to be seen. And while all four nations have agreed to work within NATO frameworks, Finland and Sweden have yet to gain membership.

[…]

Source: Nordic Air Defense Pact Combines Forces Of Hundreds Of Fighter Aircraft

13-Sided Shape That never repeats discovered

Posted on by

Computer scientists found the holy grail of tiles. They call it the “einstein,” one shape that alone can cover a plane without ever repeating a pattern.

And all it takes for this special shape is 13 sides.

In the world of mathematics, an “aperiodic monotile”—also known as an einstein based off a German phrase for one stone—is a shape that can tile a plane, but never repeat.

“In this paper we present the first true aperiodic monotile, a shape that forces aperiodicity through geometry alone, with no additional constrains applied via matching conditions,” writes Craig Kaplan, a computer science professor from the University of Waterloo and one of the four authors of the paper. “We prove that this shape, a polykite that we call ‘the hat,’ must assemble into tilings based on a substitution system.”

[…]

The history of the aperiodic tile has never had a breakthrough like this one. The first aperiodic sets had over 20,000 tiles, Kaplan tweets. “Subsequent research lowered that number, to sets of size 92, then six, and then two in the form of the famous Penrose tiles.” But those Penrose tiles were from 1974.

[…]

The team proved the nature of the shape through computer coding, and in a fascinating aside, the shape doesn’t lose its aperiodic nature even when the length of sides changes.

“We finally,” Kaplan says, “got down to one!”

It’s time for that bathroom remodel.

Source: Researchers Discovered a New 13-Sided Shape

Gen-2 by Runway text to Video AI

Posted on by

No lights. No camera. All action.Realistically and consistently synthesize new videos. Either by applying the composition and style of an image or text prompt to the structure of a source video (Video to Video). Or, using nothing but words (Text to Video). It’s like filming something new, without filming anything at all.

Visit the page for examples

Source: Gen-2 by Runway

Runway also provided Stable Diffusion, the picture generator

GitHub Copilot now integrates way better into Visual Studio (?=.* Code)

Posted on by

[…] Introduced last summer after a year-long technical trial, Copilot offers coding suggestions, though not always good ones, to developers using GitHub with supported text editors and IDEs, like Visual Studio Code.

As of last month, according to GitHub, Copilot had a hand in 46 percent of the code being created on Microsoft’s cloud repo depot and had helped developers program up to 55 percent faster.

On Wednesday, Copilot – an AI “pair programmer”, as GitHub puts it – will be ready to converse with developers ChatGPT-style in either Visual Studio Code or Visual Studio. Prompt-and-response conversations take place in an IDE sidebar chat window, as opposed to the autocompletion responses that get generated from comment-based queries in a source file.

“Copilot chat is not just a chat window,” said Dohmke. “It recognizes what code a developer has typed, what error messages are shown, and it’s deeply embedded into the IDE.”

A developer thus can highlight, say, a regex in a source file and invite Copilot to explain what the obtuse pattern matching expression does. Copilot can also be asked to generate tests, to analyze and debug, to propose a fix, or to attempt a custom task. The model can even add comments that explain source code and can clean files up like a linter.

More interesting still, Copilot can be addressed by voice. Using spoken prompts, the assistive software can produce (or reproduce) code and run it on demand. It’s a worthy accessibility option at least.

[…]

When making a pull request under the watchful eye of AI, developers can expect to find GitHub’s model will fill out tags that serve to provide additional information about what’s going on. It then falls to developers to accept or revise the suggestions.

[…]

What’s more, Copilot’s ambit has been extended to documentation. Starting with documentation for React, Azure Docs, and MDN, developers can pose questions and get AI-generated answers through a chat interface. In time, according to Dohmke, the ability to interact with documentation via a chat interface will be extended to any organization’s repositories and internal documentation.

[…]

GitHub has even helped Copilot colonize the command line, with GitHub Copilot CLI. If you’ve ever forgotten an obscure command line incantation or command flag, Copilot has you covered

[…]

Source: GitHub Copilot has some new tricks up its sleeve • The Register

EU right to repair law could see fixes for up to 10 years for more goods, still offers ways out though

Posted on by

The European Commission has adopted a new set of right to repair rules that, among other things, will add electronic devices like smartphones and tablets to a list of goods that must be built with repairability in mind.

The new rules [PDF] will need to be need to be negotiated between the European Parliament and member states before they can be turned into law. If they are, a lot more than just repairability requirements will change.

One provision will require companies selling consumer goods in the EU to offer repairs (as opposed to just replacing a damaged device) free of charge within a legal guarantee period unless it would be cheaper to replace a damaged item.

Note: so any company can get out of it quite easily.

Beyond that, the directive also adds a set of rights for device repairability outside of legal guarantee periods that the EC said will help make repair a better option than simply tossing a damaged product away.

Under the new post-guarantee period rule, companies that produce goods the EU defines as subject to repairability requirements (eg, appliances, commercial computer hardware, and soon cellphones and tablets) are obliged to repair such items for five to 10 years after purchase if a customer demands so, and the repair is possible.

[…]

The post-guarantee period repair rule also establishes the creation of an online “repair matchmaking platform” for EU consumers, and calls for the creation of a European repair standard that will “help consumers identify repairers who commit to a higher quality.”

[…]

New rules don’t do enough, say right to repair advocates

The Right to Repair coalition said in a statement that, while it welcomes the step forward taken by the EU’s new repairability rules, “the opportunity to make the right to repair universal is missed.”

While the EC’s rules focus on cutting down on waste by making products more easily repairable, they don’t do anything to address repair affordability or anti-repair practices, R2R said. Spare parts and repair charges, the group argues, could still be exorbitantly priced and inaccessible to the average consumer.

[…]

Ganapini said that truly universal right to repair laws would include assurances that independent providers were available to conduct repairs, and that components, manuals and diagnostic tools would be affordably priced. She also said that, even with the addition of smartphones and tablets to repairability requirements, the products it applies to is still too narrow.

[…]

Source: EU right to repair law could see fixes for up to 10 years • The Register

“Click-to-cancel” rule would penalize companies that make you cancel by phone

Posted on by

Canceling a subscription should be just as easy as signing up for the service, the Federal Trade Commission said in a proposed “click-to-cancel” rule announced today. If approved, the plan “would put an end to companies requiring you to call customer service to cancel an account that you opened on their website,” FTC commissioners said.

[…]

The FTC said the proposed rule would be enforced with civil penalties and let the commission return money to harmed consumers.

“The proposal states that if consumers can sign up for subscriptions online, they should be able to cancel online, with the same number of steps. If consumers can open an account over the phone, they should be able to cancel it over the phone, without endless delays,” FTC Chair Lina Khan wrote.

[…]

Source: “Click-to-cancel” rule would penalize companies that make you cancel by phone | Ars Technica

We need this globally!

Is a penguin heavy? New study explores why we disagree so often

Posted on by

Is a dog more similar to a chicken or an eagle? Is a penguin noisy? Is a whale friendly?

Psychologists at the University of California, Berkeley, say these absurd-sounding questions might help us better understand what’s at the heart of some of society’s most vexing arguments.

Research published online Thursday in the journal Open Mind shows that our concepts about and associations with even the most basic words vary widely. At the same time, people tend to significantly overestimate how many others hold the same conceptual beliefs — the mental groupings we create as shortcuts for understanding similar objects, words or events.

It’s a mismatch that researchers say gets at the heart of the most heated debates, from the courtroom to the dinner table.

“The results offer an explanation for why people talk past each other,” said Celeste Kidd, an assistant professor of psychology at UC Berkeley and the study’s principal investigator. “When people are disagreeing, it may not always be about what they think it is. It could be stemming from something as simple as their concepts not being aligned.”

Simple questions like, “What do you mean?” can go a long way in preventing a dispute from going off the rails, Kidd said. In other words, she said, “Just hash it out.”

[…]

But measuring just how much those concepts vary is a long-standing mystery.

To help understand it a bit better, Kidd’s team recruited more than 2,700 participants for a two-phase project. Participants in the first phase were divided in half and asked to make similarity judgements about whether one animal — a finch, for example — was more similar to one of two other animals, like a whale or a penguin. The other half were asked to make similarity judgments about U.S. politicians, including George W. Bush, Donald Trump, Hillary Clinton and Joe Biden. The researchers chose those two categories because people are more likely to view common animals similarly; they’d have more shared concepts. Politicians, on the other hand, might generate more variability, since people have distinct political beliefs.

But they found significant variability in how people conceptualized even basic animals.

Take penguins. The probability that two people selected at random will share the same concept about penguins is around 12%, Kidd said. That’s because people are disagreeing about whether penguins are heavy, presumably because they haven’t lifted a penguin.

“If people’s concepts are totally aligned, then all of those similarity judgments should be the same,” Kidd said. “If there’s variability in those judgments, that tells us that there’s something compositionally that’s different.”

Researchers also asked participants to guess what percentage of people would agree with their individual responses. Participants tended to believe — often incorrectly — that roughly two-thirds of the population would agree with them. In some examples, participants believed they were in the majority, even when essentially nobody else agreed with them.

It’s a finding befitting of a society of people convinced they’re right, when they’re actually wrong.

Overall, two people picked at random during the study timeframe of 2019-2021 were just as likely to have agreed as disagreed with their answers. And, perhaps unsurprisingly in a polarized society, political words were far less likely to have a single meaning — there was more disagreement — than animal words.

[…]

In a second phase of the project, participants listed 10 single-word adjectives to describe the animals and the politicians. Participants then rated the animals’ and politicians’ features — “Is a finch smart?” was an example of a question they were asked.

Again, researchers found that people differed radically in how they defined basic concepts, like about animals. Most agreed that seals are not feathered, but are slippery. However, they disagreed about whether seals are graceful. And while most people were in agreement that Trump is not humble and is rich, there was significant disagreement about whether he is interesting.

This research is significant, Kidd said, because it further shows how most people we meet will not have the exact same concept of ostensibly clear-cut things, like animals. Their concepts might actually be radically different from each other. The research transcends semantic arguments, too. It could help track how public perceptions of major public policies evolve over time and whether there’s more alignment in concepts or less.

“When people are disagreeing, it may not always be about what they think it is,” Kidd said. “It could be stemming from something as simple as their concepts not being aligned.”

Source: I say dog, you say chicken? New study explores why we disagree so often | Berkeley News

Planting Undetectable Backdoors in Machine Learning Models

Posted on by

[…]

We show how a malicious learner can plant an undetectable backdoor into a classifier. On the surface, such a backdoored classifier behaves normally, but in reality, the learner maintains a mechanism for changing the classification of any input, with only a slight perturbation. Importantly, without the appropriate “backdoor key,” the mechanism is hidden and cannot be detected by any computationally-bounded observer. We demonstrate two frameworks for planting undetectable backdoors, with incomparable guarantees.•First, we show how to plant a backdoor in any model, using digital signature schemes. The construction guarantees that given query access to the original model and the backdoored version, it is computationally infeasible to find even a single input where they differ. This property implies that the backdoored model has generalization error comparable with the original model. Moreover, even if the distinguisher can request backdoored inputs of its choice, they cannot backdoor a new input—a property we call non-replicability.•Second, we demonstrate how to insert undetectable backdoors in models trained using the Random Fourier Features (RFF) learning paradigm (Rahimi, Recht; NeurIPS 2007). In this construction, undetectability holds against powerful white-box distinguishers: given a complete description of the network and the training data, no efficient distinguisher can guess whether the model is “clean” or contains a backdoor.

[…]

Our construction of undetectable backdoors also sheds light on the related issue of robustness to adversarial examples. In particular, by constructing undetectable backdoor for an “adversarially-robust” learning algorithm, we can produce a classifier that is indistinguishable from a robust classifier, but where every input has an adversarial example! In this way, the existence of undetectable backdoors represent a significant theoretical roadblock to certifying adversarial robustness.

Source: Planting Undetectable Backdoors in Machine Learning Models : [Extended Abstract] | IEEE Conference Publication | IEEE Xplore

Whistleblowers Take Note: Don’t Trust Cropping Tools – you can often uncrop them

Posted on by

[…] It is, in fact, possible to uncrop images and documents across a variety of work-related computer apps. Among the suites that include the ability are Google Workspace, Microsoft Office, and Adobe Acrobat.

Being able to uncrop images and documents poses risks for sources who may be under the impression that cropped materials don’t contain the original uncropped content.

One of the hazards lies in the fact that, for some of the programs, downstream crop reversals are possible for viewers or readers of the document, not just the file’s creators or editors. Official instruction manuals, help pages, and promotional materials may mention that cropping is reversible, but this documentation at times fails to note that these operations are reversible by any viewers of a given image or document.

For instance, while Google’s help page mentions that a cropped image may be reset to its original form, the instructions are addressed to the document owner. “If you want to undo the changes you’ve made to your photo,” the help page says, “reset an image back to its original photo.” The page doesn’t specify that if a reader is viewing a Google Doc someone else created and wants to undo the changes the editor made to a photo, the reader, too, can reset the image without having edit permissions for the document.

For users with viewer-only access permissions, right-clicking on an image doesn’t yield the option to “reset image.” In this situation, however, all one has to do is right-click on the image, select copy, and then paste the image into a new Google Doc. Right-clicking the pasted image in the new document will allow the reader to select “reset image.” (I’ve put together an example to show how the crop reversal works in this case.)

[…]

Uncropped versions of images can be preserved not just in Office apps, but also in a file’s own metadata. A photograph taken with a modern digital camera contains all types of metadata. Many image files record text-based metadata such as the camera make and model or the GPS coordinates at which the image was captured. Some photos also include binary data such as a thumbnail version of the original photo that may persist in the file’s metadata even after the photo has been edited in an image editor.

Images and photos are not the only digital files susceptible to uncropping: Some digital documents may also be uncropped. While Adobe Acrobat has a page-cropping tool, the instructions point out that “information is merely hidden, not discarded.” By manually setting the margins to zero, it is possible to restore previously cropped areas in a PDF file.

[…]

Images and documents should be thoroughly stripped of metadata using tools such as ExifTool and Dangerzone. Additionally, sensitive materials should not be edited through online tools, as the potential always exists for original copies of the uploaded materials to be preserved and revealed.

[…]

 

Source: Whistleblowers Take Note: Don’t Trust Cropping Tools

Amazon Warns Staff Not to Share Confidential Information With ChatGPT

Posted on by

[…]

Soon, an Amazon corporate lawyer chimed in. She warned employees not to provide ChatGPT with “any Amazon confidential information (including Amazon code you are working on),” according to a screenshot of the message seen by Insider.

The attorney, a senior corporate counsel at Amazon, suggested employees follow the company’s existing conflict of interest and confidentiality policies because there have been “instances” of ChatGPT responses looking similar to internal Amazon data.

“This is important because your inputs may be used as training data for a further iteration of ChatGPT, and we wouldn’t want its output to include or resemble our confidential information (and I’ve already seen instances where its output closely matches existing material),” the lawyer wrote.

[…]

“OpenAI is far from transparent about how they use the data, but if it’s being folded into training data, I would expect corporations to wonder: After a few months of widespread use of ChatGPT, will it become possible to extract private corporate information with cleverly crafted prompts?” said Emily Bender, who teaches computational linguistics at University of Washington.

[…]

some Amazonians are already using the AI tool as a software “coding assistant” by asking it to improve internal lines of code, according to Slack messages seen by Insider.

[…]

For Amazon employees, data privacy seems to be the least of their concerns. They said using the chatbot at work has led to “10x in productivity,” and many expressed a desire to join internal teams developing similar services.

[…]

Source: Amazon Warns Staff Not to Share Confidential Information With ChatGPT

A persistent influence of supernovae on biodiversity

Posted on by

The number of exploding stars (supernovae) has significantly influenced marine life’s biodiversity during the last 500 million years. This is the essence of a new study published in Ecology and Evolution by Henrik Svensmark, DTU space.

 

Extensive studies of the fossil record have shown that the diversity of life forms has varied significantly over , and a fundamental question of evolutionary biology is which processes are responsible for these variations.

The new study reveals a major surprise: The varying number of nearby exploding stars (supernovae) closely follows changes in marine genera (the taxonomic rank above species) biodiversity during the last 500 million years. The agreement appears after normalizing the marine diversity curve by the changes in shallow marine areas along the continental coasts.

Shallow marine shelves are relevant since most lives in these areas, and changes in shelf areas open new regions where species can evolve. Therefore, changes in available shallow areas influence biodiversity.

“A possible explanation for the supernova-diversity link is that supernovae influence Earth’s climate,” says Henrik Svensmark, author of the paper and senior researcher at DTU Space.

“A high number of supernovae leads to a with a large temperature difference between the equator and polar regions. This results in stronger winds, ocean mixing, and transportation of life-essential nutrients to the along the continental shelves.”

Variations in relative supernova history (black curve) compared with genera-level diversity curves normalized with the area of shallow marine margins (shallow areas along the coasts). The brown and light green curves are major marine animals’ genera-level diversity. The orange is marine invertebrate genera-level diversity. Finally, the dark green curve is all marine animals’ genera-level diversity. Abbreviations for geological periods are Cm Cambrian, O Ordovician, S Silurian, D Devonian, C Carboniferous, P Permian, Tr Triassic, J Jurassic, K Cretaceous, Pg Palaeogene, Ng Neogene. Credit: Henrik Svensmark, DTU Space

The paper concludes that supernovae are vital for primary bioproductivity by influencing the transport of nutrients. Gross primary bioproductivity provides energy to the , and speculations have suggested that changes in bioproductivity may influence biodiversity. The present results are in agreement with this hypothesis.

“The new evidence points to a connection between life on Earth and supernovae, mediated by the effect of cosmic rays on clouds and climate,” says Henrik Svensmark.

When heavy stars explode, they produce cosmic rays, which are elementary particles with enormous energies. Cosmic rays travel to our solar system, where some end their journey by colliding with Earth’s atmosphere. Previous studies by Henrik Svensmark and colleagues referenced below show that they become the primary source of ions help form and grow aerosols required in cloud formation.

Since clouds can regulate the solar energy reaching Earth’s surface, the cosmic-ray-aerosol-cloud influences climate. Evidence shows substantial climate shifts when the intensity of changes by several hundred percent over millions of years.

More information: Henrik Svensmark, A persistent influence of supernovae on biodiversity over the Phanerozoic, Ecology and Evolution (2023). DOI: 10.1002/ece3.9898

Henrik Svensmark, Supernova Rates and Burial of Organic Matter, Geophysical Research Letters (2022). DOI: 10.1029/2021GL096376

Svensmark, H. and Friis-Christensen, E., Variation of Cosmic Ray Flux and Global Cloud Coverage -A missing Link in Solar-Climate Relationships, Journal of Atmospheric and Terrestrial Physics, 59, 1225, (1997)

Nir J. Shaviv et al, The Phanerozoic climate, Annals of the New York Academy of Sciences (2022). DOI: 10.1111/nyas.14920

Henrik Svensmark, Evidence of nearby supernovae affecting life on Earth, Monthly Notices of the Royal Astronomical Society (2012). DOI: 10.1111/j.1365-2966.2012.20953.x

Source: A persistent influence of supernovae on biodiversity

Ubisoft Ghostwriter: AI to write NPC dialogue

Posted on by

[…] As games grow bigger in scope, writers are facing the ratcheting challenge of keeping NPCs individually interesting and realistic. How do you keep each interaction with them – especially if there are hundreds of them – distinct? This is where Ghostwriter, an in-house AI tool created by Ubisoft’s R&D department, La Forge, comes in.

Ghostwriter isn’t replacing the video game writer, but instead, alleviating one of the video game writer’s most laborious tasks: writing barks. Ghostwriter effectively generates first drafts of barks – phrases or sounds made by NPCs during a triggered event – which gives scriptwriters more time to polish the narrative elsewhere. Ben Swanson, R&D Scientist at La Forge Montreal, is the creator of Ghostwriter, and remembers the early seeds of it ahead of his presentation of the tech at GDC this year.

[…]

Ghostwriter is the result of conversations with narrative designers who revealed a challenge, one that Ben identified could be solved with an AI tool. Crowd chatter and barks are central features of player immersion in games – NPCs speaking to each other, enemy dialogue during combat, or an exchange triggered when entering an area all provide a more realistic world experience and make the player feel like the game around them exists outside of their actions. However, both require time and creative effort from scriptwriters that could be spent on other core plot items. Ghostwriter frees up that time, but still allows the scriptwriters a degree of creative control.

“Rather than writing first draft versions themselves, Ghostwriter lets scriptwriters select and polish the samples generated,” Ben explains. This way, the tech is a tool used by the teams to support them in their creative journey, with every interaction and feedback originating from the members who use it.

As a summary of its process, scriptwriters first create a character and a type of interaction or utterance they would like to generate. Ghostwriter then proposes a select number of variations which the scriptwriter can then choose and edit freely to fit their needs. This process uses pairwise comparison as a method of evaluation and improvement. This means that, for each variation generated, Ghostwriter provides two choices which will be compared and chosen by the scriptwriter. Once one is selected, the tool learns from the preferred choice and, after thousands of selections made by humans, it becomes more effective and accurate.

[…]

The team’s ambition is to give this AI power to narrative designers, who will be able to eventually create their own AI system themselves, tailored to their own design needs. To do this, they created a user-friendly back-end tool website called Ernestine, which allows anyone to create their own machine learning models used in Ghostwriter. Their hope is that teams consider Ghostwriter before they start their narrative process and create their models with a vision in mind, effectively making the tech an integral part of the production pipeline.

[…]

Source: The Convergence of AI and Creativity: Introducing Ghostwriter

This looks like another excellent way of employing generative AI in a way that eases the life of people doing shitty jobs.