Follow Slashdot stories on Twitter

 


Forgot your password?
Close
typodupeerror
×

Submission + - LAION-5B Dataset Removed After Discovery of Child Sexual Abuse Material (404media.co)

Submitted by samleecole
samleecole writes: The LAION-5B machine learning dataset used by Google, Stable Diffusion, and other major AI products has been removed by the organization that created it after a Stanford study found that it contained 3,226 suspected instances of child sexual abuse material, 1,008 of which were externally validated.

LAION told 404 Media on Tuesday that out of “an abundance of caution,” it was taking down its datasets temporarily “to ensure they are safe before republishing them."

According to a new study by the Stanford Internet Observatory shared with 404 Media ahead of publication, the researchers found the suspected instances of CSAM through a combination of perceptual and cryptographic hash-based detection and analysis of the images themselves.

“We find that having possession of a LAION5B dataset populated even in late 2023 implies the possession of thousands of illegal images—not including all of the intimate imagery published and gathered nonconsensually, the legality of which is more variable by jurisdiction,” the paper says. “While the amount of CSAM present does not necessarily indicate that the presence of CSAM drastically influences the output of the model above and beyond the model’s ability to combine the concepts of sexual activity and children, it likely does still exert influence. The presence of repeated identical instances of CSAM is also problematic, particularly due to its reinforcement of images of specific victims.”

The finding highlights the danger of largely indiscriminate scraping of the internet for the purposes of generative artificial intelligence.

Submission + - Attack discovered against SSH (arstechnica.com) 1

Submitted by jd
jd writes: Ars Technica is reporting a newly-discovered man-in-the-middle attack against SSH. This only works if you are using "ChaCha20-Poly1305" or "CBC with Encrypt-then-MAC", so it isn't a universal flaw. The CVE numbers for this vulnerability are CVE-2023-48795, CVE-2023-46445, and CVE-2023-46446.

From TFA:

At its core, Terrapin works by altering or corrupting information transmitted in the SSH data stream during the handshake—the earliest stage of a connection, when the two parties negotiate the encryption parameters they will use to establish a secure connection. The attack targets the BPP, short for Binary Packet Protocol, which is designed to ensure that adversaries with an active position can't add or drop messages exchanged during the handshake. Terrapin relies on prefix truncation, a class of attack that removes specific messages at the very beginning of a data stream.

The Terrapin attack is a novel cryptographic attack targeting the integrity of the SSH protocol, the first-ever practical attack of its kind, and one of the very few attacks against SSH at all. The attack exploits weaknesses in the specification of SSH paired with widespread algorithms, namely ChaCha20-Poly1305 and CBC-EtM, to remove an arbitrary number of protected messages at the beginning of the secure channel, thus breaking integrity. In practice, the attack can be used to impede the negotiation of certain security-relevant protocol extensions. Moreover, Terrapin enables more advanced exploitation techniques when combined with particular implementation flaws, leading to a total loss of confidentiality and integrity in the worst case.

Submission + - Blue Origin's Suborbital Rocket Flies For First Time In 15 Months (arstechnica.com)

Submitted by Anonymous Coward
An anonymous reader writes: With redesigned engine components, Blue Origin's New Shepard rocket took off from West Texas and flew to the edge of space on Tuesday with a package of scientific research and technology demonstration experiments. This was the first flight of Blue Origin's 60-foot-tall (18-meter) New Shepard rocket since September 12, 2022, when an engine failure destroyed the booster and triggered an in-flight abort for the vehicle's pressurized capsule. There were no passengers aboard for that mission, and the capsule safely separated from the failed booster and parachuted to a controlled landing.

The flight on Tuesday also didn't carry people. Instead, Blue Origin, Jeff Bezos's space company, lofted 33 payloads from NASA, research institutions, and commercial companies. Some of these payloads were flown again on Tuesday's launch after failing to reach space on the failed New Shepard mission last year. Among these payloads were an experiment to demonstrate hydrogen fuel cell technology in microgravity and an investigation studying the strength of planetary soils under different gravity conditions. Blue Origin's capsule, mounted on top of the rocket, also flew 38,000 postcards submitted by students through Club for the Future, the company's nonprofit.

For Tuesday's return-to-flight mission, the New Shepard rocket ignited its BE-3PM engine and climbed away from Blue Origin's remote launch site near Van Horn, Texas, at 10:42 am CST (16:42 UTC). The hydrogen-fueled engine fired for more than two minutes, then shut down as scheduled as the rocket continued coasting upward, reaching an altitude of more than 347,000 feet (106 kilometers). The booster returned for a precision propulsive landing a short distance from the launch pad, and Blue Origin's capsule deployed three parachutes to settle onto the desert floor, completing a 10-minute up-and-down flight. Blue Origin has launched 24 missions with its reusable New Shepard rocket, including six flights carrying people just over the Kármán line, the internationally recognized boundary of space 100 kilometers above Earth.

Submission + - TomTom Creates AI-Based Conversational Assistant For Vehicles With Microsoft (reuters.com)

Submitted by Anonymous Coward
An anonymous reader writes: Digital mapping specialist TomTom said on Tuesday it has partnered with tech giant Microsoft to create an artificial intelligence (AI)-powered conversational assistant for vehicles. The assistant will allow users to "converse naturally with their vehicles" and enable voice interaction with infotainment, location search, and vehicle command systems, the company said.

TomTom, which competes with Google Maps and the world's biggest mapping platform HERE, used various Microsoft services like its Azure OpenAI Service to create the voice assistant. The Microsoft Azure OpenAI Service allows enterprises to leverage ChatGPT maker OpenAI's large language models (LLM). The voice assistant can be integrated into other automotive infotainment systems and is also built into TomTom's Digital Cockpit, an open, modular in-vehicle infotainment platform, the Dutch map maker said. The company began working with Microsoft in 2016, when it first started powering Azure Maps location services.

Submission + - Meta's News Ban In Canada Remains As Online News Act Goes Into Effect (bbc.com)

Submitted by Anonymous Coward
An anonymous reader writes: A bill that mandates tech giants pay news outlets for their content has come into effect in Canada amid an ongoing dispute with Facebook and Instagram owner Meta over the law. Some have hailed it as a game-changer that sets out a permanent framework that will see a steady drip of funds from wealthy tech companies to Canada's struggling journalism industry. But it has also been met with resistance by Google and Meta — the only two companies big enough to be encompassed by the law. In response, over the summer, Meta blocked access to news on Facebook and Instagram for Canadians. Google looked set to follow, but after months of talks, the federal government was able to negotiate a deal with the search giant as the company has agreed to pay Canadian news outlets $75 million annually.

No such agreement appears to be on the horizon with Meta, which has called the law "fundamentally flawed." If Meta is refusing to budge, so is the government. "We will continue to push Meta, that makes billions of dollars in profits, even though it is refusing to invest in the journalistic rigor and stability of the media," Prime Minister Justin Trudeau told reporters on Friday.

Submission + - Microsoft releases software to remove unwanted HP printer software (arstechnica.com)

Submitted by UnknowingFool
UnknowingFool writes: Microsoft has released a new software tool to remove printer software from HP that was installed without user permission or system need. A few weeks ago, users noticed that Windows update installed HP printer software even if they did not have HP printers or printers at all. Affecting Windows 10 and 11, consumers reported that this update sometimes caused problems as it could rename their non-HP printers as HP printers causing some printing features to be inaccessible. Microsoft has not disclosed the root cause of the issue.

Submission + - SPAM: Did Your Parents Give You Your Own Desk? At What Age?

Submitted by theodp
theodp writes: "To me it's obvious that they are trying to be efficient and save money but at the same time also utilize resources," Google CEO Sundar Pichai told employees in a companywide meeting last week as he defended the Cloud unit's new desk-sharing policy for employees. The company is calling the downsizing effort Cloud Office Evolution (CLOE), prompting one employee to tell Pichai that there's "no need to make every bad thing sound like a miracle."

It does seem curious that Google — whose X Development 'moonshot' R&D organization boasts "We create radical new technologies to solve some of the world’s hardest problems" — can't figure out how to provide employees with their own desks, especially considering Google's tremendous financial resources ($1.22T market cap, $1.6M revenue per employee). After all, a 1983 study found that parents managed to ensure that 98% of Japanese 10 year-olds with their own desk (compared to 63% of their American counterparts), which was viewed as indicative of parents' concern for their child's achievement.

So, did your parents provide you with your own desk? How old were you at the time?

Submission + - Fixing bufferbloat at the ISP, finally (packetpushers.net)

Submitted by mtaht
mtaht writes: Perhaps the slashdot audience would be interested in my latest attempts to finish fixing bufferbloat — now at the ISP layer — and pushing 25Gbits! across the rest of the internet, this time with a lightweight middlebox built around XDP, eBPF, Rust, and CAKE, called LibreQos, released as free software. We thought about calling our effort LibreQoE, or LibreSQM, because everyone knows QoS doesn't work, but Flow Queuing (RFC8290) and Active Queue Management do...

Submission + - Musk had Twitter engineers boost his tweets after Biden got more views (arstechnica.com) 7

Submitted by SpzToid
SpzToid writes: Twitter CEO Elon Musk had Twitter engineers set up a "special system" to boost his tweets after his post about the Super Bowl got fewer views than a tweet from President Biden, according to a report by Platformer yesterday.

Biden's tweet supporting the Philadelphia Eagles has 29.1 million views, while Musk's now-deleted tweet also supporting the Eagles received a mere 9.1 million. Hours after the Eagles' loss, Musk's cousin James Musk posted a message in Twitter's Slack asking anyone "who can make dashboards and write software" to help with "debugging an issue with engagement across the platform," the report said. James Musk's Slack message reportedly called the situation one of "high urgency."

Additionally, Elon Musk "flew his private jet back to the Bay Area on Sunday night to demand answers from his team."

"Late Sunday night, Musk addressed his team in-person," Platformer wrote. "Roughly 80 people were pulled in to work on the project, which had quickly become priority number one at the company. Employees worked through the night investigating various hypotheses about why Musk's tweets weren't reaching as many people as he thought they should and testing out possible solutions."

Fix it or you’re fired

Earlier this month, Musk reportedly fired an engineer who tried to explain why his tweet views were down. More firings were threatened as Musk's "deputies told the rest of the engineering team this weekend that if the engagement issue wasn't 'fixed,' they would all lose their jobs as well," the Platformer report said.

A solution was found by Monday. Musk's newfound reach was immediately noticeable to users in the default "For You" feed that isn't limited to accounts each user follows.

...meanwhile in related news, Slashdot's shadow ban on All News or Nerdy Discussions About Elon/Twitter remains in full effect.

Submission + - SPAM: Ask Slashdot: Is Slashdot website code still open source? How to contribute?

Submitted by guest reader
guest reader writes: This Wikipedia page contains the following note:
It is unclear whether Slashdot still uses Slash or some newer look-alike since around 2009.

The wikipedia page continues:
Slashdot runs on Slash, a content management system available under the GNU General Public License. Early versions of Slash were written by Rob Malda in the spring of 1998. After Andover.net bought Slashdot in June 1999, several programmers were hired to structure the code and render it scalable, as its users had increased from a few hundred to tens of thousands. This work was done by Brian Aker, Patrick Galbraith and Chris Nandor, resulting in version 2 of the software, released in 2001. Slash remains Free software and anyone can contribute to development.

External links from the wiki page:
- Slash on SouceForge which has last update 2016-03-11
- slashcode on GitHub which has last update 2015-04-21
- rehash on GitHub, SoylentNews Rehash code which has last update 2023-01-01

About Slashdot contains the following information:
This is Slashdot, a website based on and running the Slashdot-Like Automated Story-Telling Homepage software. This link then points to Slash Git which has last update 2009-09-22.
Link to Original Source

Submission + - $120bn wiped off Google after Bard AI chatbot gives wrong answer (telegraph.co.uk)

Submitted by Anonymous Coward
An anonymous reader writes: Company's answer to Microsoft's ChatGPT AI falls at the first hurdle

More than $120bn was wiped off Google’s market value after its new AI search assistant gave a wrong answer that was featured in promotional material.

Parent company Alphabet’s share price dropped 8pc after the Google Bard tool’s launch got off to a rocky start following its misleading response to a question about a NASA telescope.

At its lowest point on Wednesday Alphabet was trading at $98.08 (£81.16), a fall of 8.1pc on the previous day’s price of $106.77 (£88.35).

It marked the biggest one-day fall in Alphabet’s value since October 2022, when the company shed 9pc of its value in one day after unveiling a big slowdown in sales, profits and growth.

In an animated GIF showing how Bard works, a user types in the search query “what new discoveries from the James Webb Space Telescope can I tell my 9 year old about?”

The NASA telescope was made operational in December 2021 and has been used by scientists to make several discoveries of new planets outside the Solar System.

One of the responses generated by Bard says: “JWST took the very first pictures of a planet outside of our own solar system.”

This is not accurate. The first picture ever taken of a planet outside the solar system – an exoplanet – was captured in 2004 by the Very Large Telescope array in Chile.

Submission + - UK court: Craig Wright cannot copyright Bitcoin file format (decrypt.co)

Submitted by UnknowingFool
UnknowingFool writes: UK Judge James Mellor has thrown out Craig Wright's cases against Bitcoin derivatives like Bitcoin Cash as Wright cannot claim copyright on the Bitcoin file format. Wright had sued forks of Bitcoin claiming they breached his copyrights to prevent them from operating. The judge disagreed noting that Wright had failed to meet a requirement of copyright called "fixation" detailing where/when/how the original expression was first recorded somewhere in any media.

“Whilst I accept that the law of copyright will continue to face challenges with new digital technologies, I do not see any prospect of the law as currently stated and understood in the caselaw allowing copyright protection of subject-matter which is not expressed or fixed anywhere.” wrote Judge Mellor

In other words Wright has failed to show any evidence that he wrote down the file format somewhere to claim that he created the file format. This is not the first time Wright has failed to produce credible evidence in a court case: in a Oslo, Norway case last year Wright claimed he destroyed a hard drive in 2016 containing the Nakomoto original keys despite telling a US court in 2020 that he was waiting on the same keys to be delivered by a special courier. Those keys were later ruled to be fictitious.

Submission + - Rackspace hosted exchange down for four days (and counting) (rackspace.com)

Submitted by neilo_1701D
neilo_1701D writes: Sometime early Friday morning, Rackspace hosted Exchange stopped working. Late Friday night, they reported the outage as a "security event"; since then, updates have been sporadic and suggestions to get email flowing again convoluted. They first started offering Microsoft-hosted mailboxes, and as that solution seemed to go down in flames they offered to redirect mail to externals accounts (gmail etc).

In the meantime, hosted email remains offline and with no timeline for restoration and increasingly desperate customers look for non-Rackspace solutions. Companies (and individuals) with decades of email history risk losing that archive with all the options that Rackspace offer.

Slashdot Top Deals

Any sufficiently advanced technology is indistinguishable from a rigged demo. - Andy Finkel, computer guy

Close