Linux fréttir

Lands a fourth giant customer for bespoke AI accelerators

Broadcom has impressed investors by posting record revenues but admitted it has sold a lot of shelfware to VMware customers.…

An anonymous reader quotes a report from The Guardian: Fine-particulate air pollution can drive devastating forms of dementia by triggering the formation of toxic clumps of protein that destroy nerve cells as they spread through the brain, research suggests. Exposure to the airborne particles causes proteins in the brain to misfold into the clumps, which are hallmarks of Lewy body dementia, the second most common form of dementia after Alzheimer's disease. The finding has "profound implications" for preventing the neurodegenerative disorder, which affects millions worldwide, with scientists calling for a concerted effort to improve air quality by cutting emissions from industrial activity and vehicle exhausts, improving wildfire management and reducing wood burning in homes. The researchers began by analyzing hospital records of the 56.5 million US Medicare patients. They looked at those who were admitted for the first time between 2000 and 2014 with the protein damage. Armed with the patients' zip codes, the scientists estimated their long-term exposure to PM2.5 pollution, airborne particles that are smaller than 2.5 thousandths of a millimeter. These can be inhaled deep into the lungs and are found in the bloodstream, brain and other organs. They found that long-term exposure to PM2.5 raised the risk of Lewy body dementia, but had less of an impact on rates of another neurodegenerative brain disease that is not driven by the toxic proteins. Lewy bodies are made from a protein called alpha-synuclein. The protein is crucial for healthy brain functioning, but can misfold in various ways to produce different kinds of harmful Lewy bodies. These can kill nerve cells and cause devastating disease by spreading through the brain. To see if air pollution could trigger Lewy bodies, the team exposed mice to PM2.5 pollution every other day for 10 months. Some were normal mice, but others were genetically modified to prevent them making alpha-synuclein. The results were striking: in normal mice, nerve cells died off, leading to brain shrinkage and cognitive decline. The genetically modified mice were largely unaffected. Further work in mice showed that PM2.5 pollution drove the formation of aggressive, resilient and toxic clumps of alpha-synuclein clumps that looked very similar to Lewy bodies in humans. Although the work is in mice, the findings are considered compelling evidence. The work has been published in the journal Science.

Read more of this story at Slashdot.

Rideshare giant also reveals 350-petabyte data lake it protects with tech adapted from Airbnb

Uber’s Indian arm has started using its app to offer rideshare and delivery drivers the chance to make a Rupee by classifying data used by AI systems.…

In a landmark collaboration across 22 labs, neuroscientists have created the first brain-wide map of decision-making in mice, tracking over 620,000 neurons across nearly 280 brain regions. They found that decision-making is distributed much more broadly than previously thought, involving not just "cognitive" centers but also regions linked to movement. From a report: The task was deceptively simple task. Mice sat in front of a screen that intermittently displayed a black-and-white striped circle for a brief amount of time on either the left or right side. A mouse could earn a sip of sugar water if they quickly moved the circle toward the center of the screen by operating a tiny steering wheel in the same direction, often doing so within one second. On some trials, the circle was faint, requiring the animal to rely on past experience to make a guess, which allowed researchers to study how expectations influence future decisions. While the mice performed the task, researchers recorded brain activity using high-density electrodes that allowed them to monitor hundreds of neurons across many regions simultaneously. The work was divided across the participating labs, so that each lab mapped a particular region of the mouse brain. The pooled dataset covers 620,000 neurons recorded from 139 mice in 12 labs, encompassing nearly the entire brain. The resulting map revealed that decision-making activity is distributed across the brain, including in areas traditionally associated with movement rather than cognition. The findings have been published in two papers in the journal Nature.

Read more of this story at Slashdot.

An Indiana bankruptcy lawyer named Mark Zuckerberg is suing Meta after his Facebook page was repeatedly shut down for "impersonating" CEO Mark Zuckerberg, despite being his real legal name. TechCrunch reports: Mark Zuckerberg the lawyer uses a commercial Facebook page to advertise his legal practice and communicate with potential clients. But his page has been disabled five times in the last eight years, since Meta's moderation systems flag his account as falsely impersonating Mark Zuckerberg, the founder of the platform. Mark Zuckerberg is not impersonating Mark Zuckerberg, because he, too, is Mark Zuckerberg. In his legal complaint, Mark Zuckerberg points out that he has been practicing law since Mark Zuckerberg was just three years old. "It's not funny," Mark Zuckerberg, the lawyer, said to Indianapolis' 13WTHR. "Not when they take my money. This really pissed me off." Mark Zuckerberg has spent over $11,000 to advertise his page on Mark Zuckerberg's Meta platforms, but when Mark Zuckerberg's account is disabled for allegedly impersonating Mark Zuckerberg, Mark Zuckerberg still has to pay for these advertisements. Zuckerberg created a website, iammarkzuckerberg.com, chronicling how his life has been shaped by being named Mark Zuckerberg. The lawsuit can be found here.

Read more of this story at Slashdot.

Warner Bros. Discovery has filed a major copyright lawsuit against Midjourney, accusing the AI image generator of exploiting its movies and TV shows to train models and generate near-identical reproductions of iconic characters like Batman, Bugs Bunny, and Rick and Morty. From The Hollywood Reporter: The company "brazenly dispenses Warner Bros. Discovery's intellectual property" by letting subscribers produce images and videos of iconic copyrighted characters, alleges the complaint, filed on Thursday in California federal court. "The heart of what we do is develop stories and characters to entertain our audiences, bringing to life the vision and passion of our creative partners," said a Warner Bros. Discovery spokesperson in a statement. "Midjourney is blatantly and purposefully infringing copyrighted works, and we filed this suit to protect our content, our partners, and our investments." For years, AI companies have been training their technology on data scraped across the internet without compensating creators. It's led to lawsuits from authors, record labels, news organizations, artists and studios, which contend that some AI tools erode demand for their content. Warner Bros. Discovery joins Disney and Universal, which earlier this year teamed up to sue Midjourney. By their thinking, the AI company is a free-rider plagiarizing their movies and TV shows. In the lawsuit, Warner Bros. Discovery points to Midjourney generating images of iconic copyrighted characters. At the forefront are heroes who're at the center of DC Studios' movies and TV shows, like Superman, Wonder Woman and The Joker; others are Looney Tunes, Tom and Jerry and Scooby-Doo characters who've become ubiquitous household names; more are Cartoon Network characters, including those from Rick and Morty, who've emerged as something of cultural touchstones in recent years. [...] The lawsuit argues Midjourney's ability to return copyrighted characters is a "clear draw for subscribers," diverting consumers away from purchasing Warner Bros. Discovery-approved posters, wall art and prints, among other products that must now compete against the service. [...] Warner Bros. Discovery seeks Midjourney's profits attributable to the alleged infringement or, alternatively, $150,000 per infringed work, which could leave the AI company on the hook for massive damages. The thrust of the studios' lawsuits will likely be decided by one question: Are AI companies covered by fair use, the legal doctrine in intellectual property law that allows creators to build upon copyrighted works without a license? The lawsuit can be found here.

Read more of this story at Slashdot.