news aggregator

A Tanzanian entrepreneur is transforming cement bags into solar-powered backpacks, helping students study after dark in areas without electricity. Innocent James's company, Soma Bags, sold 36,000 solar backpacks across Africa last year, with prices ranging from 12,000 to 22,500 Tanzanian shillings ($4-8), according to CNN. The innovation comes as 600 million Africans lack electricity access. In Tanzania, fewer than half of mainland households have power, forcing families to rely on expensive kerosene lamps. The backpacks, manufactured in James's Bulale factory employing 65 staff, feature flexible solar panels that charge during students' walks to school. One day of sunlight provides six to eight hours of reading light, making them more cost-effective than kerosene lamps commonly used in Tanzania, where fewer than half of mainland households have electricity access.

Read more of this story at Slashdot.

Bruce Schneier and Nathan E. Sanders, writing in a post: Someone who makes calculus mistakes is also likely to respond "I don't know" to calculus-related questions. To the extent that AI systems make these human-like mistakes, we can bring all of our mistake-correcting systems to bear on their output. But the current crop of AI models -- particularly LLMs -- make mistakes differently. AI errors come at seemingly random times, without any clustering around particular topics. LLM mistakes tend to be more evenly distributed through the knowledge space. A model might be equally likely to make a mistake on a calculus question as it is to propose that cabbages eat goats. And AI mistakes aren't accompanied by ignorance. A LLM will be just as confident when saying something completely wrong -- and obviously so, to a human -- as it will be when saying something true. The seemingly random inconsistency of LLMs makes it hard to trust their reasoning in complex, multi-step problems. If you want to use an AI model to help with a business problem, it's not enough to see that it understands what factors make a product profitable; you need to be sure it won't forget what money is. [...] Humans may occasionally make seemingly random, incomprehensible, and inconsistent mistakes, but such occurrences are rare and often indicative of more serious problems. We also tend not to put people exhibiting these behaviors in decision-making positions. Likewise, we should confine AI decision-making systems to applications that suit their actual abilities -- while keeping the potential ramifications of their mistakes firmly in mind.

Read more of this story at Slashdot.

Big organizations and governments are main users of these gateways

SonicWall is warning customers of a critical vulnerability that was potentially already exploited as a zero-day.…

Britain's competition watchdog launched investigations into Apple and Google's mobile ecosystems on Thursday under new powers to tackle digital market abuses that took effect this year. The Competition and Markets Authority will examine whether the tech giants' control over operating systems, app stores and browsers constitutes "strategic market status" requiring regulatory intervention. The probe will focus on potential barriers to competition, preferential treatment of their own apps, and whether developers face unfair terms for app distribution. The regulator could force changes including mandatory access to key mobile functions or allowing users to download apps outside official stores.

Read more of this story at Slashdot.

Company 'strongly disagrees' with law infringement allegations

Meta has again come under fire for its pay-or-consent model in the EU.…

People with attention deficit hyperactivity disorder face significantly shorter life expectancy and higher mental health risks, a British study of over 30,000 patients found. The research, published in The British Journal of Psychiatry, showed men with ADHD lived 4.5 to 9 years less, while women's lives were shortened by 6.5 to 11 years. The study compared primary care data from 30,029 adults with ADHD against 300,400 people without the condition. "Although many people with ADHD live long and healthy lives, our finding that on average they are living shorter lives than they should indicates unmet support needs," said Dr. Liz O'Nions, honorary research fellow at University College London. The study linked ADHD to increased risks of anxiety, depression, self-harm, and suicide, along with higher rates of smoking and alcohol use.

Read more of this story at Slashdot.

Generative AI needs a break, just like the rest of us, m'kay?

OUTAGE Reactivate your brain. ChatGPT has gone down.…

Pakistan's lower house of parliament on Thursday passed a controversial bill that will give the government sweeping controls on social media, including sending users to prison for spreading disinformation. From a report: The bill was quickly passed after lawmakers from the opposition party of imprisoned former Prime Minister Imran Khan staged a walkout to denounce the law. Critics say the government is seeking to further suppress freedom of speech. Farhatullah Babar, a leading human rights activist, said the latest changes to cybercrime law were aimed at "further stifling the freedom of expression through setting up of multiple authorities under executive control, enlarging the print of unaccountable intelligence agencies." He said the law also "gives sweeping powers to the executive not only over the contents of the message but also the messengers, namely the social media platforms." Under the Prevention of Electronic Crimes Act, which was introduced in the National Assembly Wednesday, authorities would create an agency with the power to order the immediate blocking of content deemed "unlawful and offensive" from social media, such as content critical of judges, the armed forces, parliament or provincial assemblies. Individuals and organizations posting such content may also be blocked from social media.

Read more of this story at Slashdot.

Experts warn not to take leaks lightly as years-long compromises could remain undetected

Thousands of email addresses included in the Belsen Group's dump of FortiGate configs last week are now available online, revealing which organizations may have been impacted by the 2022 zero-day exploits.…

Chips and software giant Broadcom says it's 'flexible and open' on licensing terms, but customers disagree

VMware users continue to be unhappy with licensing changes since the virtualization giant was acquired by Broadcom, and are now complaining that they are being forced into three-year commitments when renewing vSphere licenses.…

Microsoft's absence from OpenAI's Stargate announcement follows months of tension between the companies and signals a new era in which the longtime partners will be less reliant on each other. From a report: At a White House press conference, the ChatGPT maker announced Stargate, a venture with Oracle and tech investor SoftBank. The new company plans to spend up to $500 billion building new data centers in the U.S. to help power OpenAI's development. The assembled leaders -- OpenAI's Sam Altman, Oracle's Larry Ellison, SoftBank's Masayoshi Son and President Trump -- discussed how AI could create jobs and even cure cancer. Microsoft CEO Satya Nadella was thousands of miles away, at the World Economic Forum in Davos, Switzerland. The developments show how the OpenAI-Microsoft partnership that helped trigger the generative-AI boom is drifting apart as each company focuses on its own evolving needs. In the months leading up to the announcement, the two sides had been haggling over what to do about OpenAI's seemingly insatiable appetite for computing power and its contention Microsoft couldn't fulfill it even though their agreement didn't allow OpenAI to easily switch to others, said people familiar with the discussions. OpenAI is almost entirely reliant on Microsoft to provide it with the data centers it needs to build and operate its sophisticated AI software. That has been part of their agreement since Microsoft first invested in 2019. With the success of ChatGPT, OpenAI's need for computing power surged. Its executives have said ending the exclusive cloud contract could be crucial to compete with rival AI developers that don't have the same constraints.

Read more of this story at Slashdot.

Shares slide at 'most profitable' company in Korea as world worries over geopolitics

Market uncertainty and fears around trade protectionism are overshadowing SK hynix's latest earnings, with its shares sliding despite revenue doubling for the financial year just completed.…

An anonymous reader quotes a report from Reuters: E-commerce giant Amazon.com is exiting its operations in the Canadian province of Quebec, leading to the loss of about 1,700 full-time jobs, the company said on Wednesday, prompting Ottawa to express its unhappiness. The online retailer will phase out operations across seven sites in the province -- the only location in Canada with unionized Amazon employees -- over the next two months. It will return to a third-party delivery model, relying on local small businesses, similar to its approach before 2020. "Following a recent review of our Quebec operations, we've seen that returning to a third-party delivery model ... will allow us to provide even more savings to our customers," Amazon spokesperson Barbara Agrait said. The move will affect approximately 250 seasonal workers. Amazon will offer affected employees a package including up to 14 weeks' pay and "transitional benefits such as job placement resources," Agrait added. "This is not the way business is done in Canada," said Federal Innovation Minister Francois-Philippe Champagne. "There is no doubt that the closings announced today are part of an anti-union campaign against CSN and Amazon employees," said CSN president Caroline Senneville in a statement. "This move contradicts the provisions of the Quebec Labour Code, which we will strongly oppose," Senneville added.

Read more of this story at Slashdot.

As Microsoft, Apple, and Google switch the tech on by default, what happened to asking for permission first?

Opinion Copilot in Microsoft 365 and Apple Intelligence on iDevices are the latest examples of the tech industry's obsession with making services opt-out rather than opt-in.…

CMA flexes its new Strategic Market Status muscles

The UK's Competition and Markets Authority (CMA) is launching Strategic Market Status (SMS) investigations into both Apple and Google, probing the duo's control of their respective mobile ecosystems.…

Coming soon to April's TrueNAS SCALE release, dubbed 'Fangtooth'

The latest version of OpenZFS offers RAID expansion, plus faster data deduplication donated by iXsystems. The code will be available very soon in the beta of TrueNAS SCALE 25.04.…

Plus: 'Largest-ever' duff traffic tsunami clocks in at 5.6 Tbps

In addition to Chinese spies invading organizations' networks and ransomware crews locking up sensitive files, botnets blasting distributed denial of service (DDoS) attacks can still cause a world of hurt — and website downtime — and it's quite likely your competitors are to blame.…

A federal district court has ruled that backdoor searches of Americans' private communications collected under Section 702 of FISA are unconstitutional without a warrant. "The landmark ruling comes in a criminal case, United States v. Hasbajrami, after more than a decade of litigation, and over four years since the Second Circuit Court of Appeals found that backdoor searches constitute 'separate Fourth Amendment events' and directed the district court to determine a warrant was required," reports the Electronic Frontier Foundation (EFF). "Now, that has been officially decreed." Longtime Slashdot reader schwit1 shares the report: Hasbajrami involves a U.S. resident who was arrested at New York JFK airport in 2011 on his way to Pakistan and charged with providing material support to terrorists. Only after his original conviction did the government explain that its case was premised in part on emails between Mr. Hasbajrami and an unnamed foreigner associated with terrorist groups, emails collected warrantless using Section 702 programs, placed in a database, then searched, again without a warrant, using terms related to Mr. Hasbajrami himself. The district court found that regardless of whether the government can lawfully warrantlessly collect communications between foreigners and Americans using Section 702, it cannot ordinarily rely on a "foreign intelligence exception" to the Fourth Amendment's warrant clause when searching these communications, as is the FBI's routine practice. And, even if such an exception did apply, the court found that the intrusion on privacy caused by reading our most sensitive communications rendered these searches "unreasonable" under the meaning of the Fourth Amendment. In 2021 alone, the FBI conducted 3.4 million warrantless searches of US person's 702 data.

Read more of this story at Slashdot.

And the government thinks that AI and taking shackles off big tech will help? God help Britain

For the first time since the start of the pandemic, the number of tech firms incorporated in the UK has declined, with a shrinking economy, as well as high inflation and interest rates causing a slump in business confidence.…

Nailed just 15% of assigned tasks

A service described as "the first AI software engineer" appears to be rather bad at its job, based on a recent evaluation.…