Tag Archives: author_name|Mariella Moon

Intel is separating its ailing foundry business from the main company

Posted on by

Intel is turning its foundry business, which manufactures chips for other companies, into an independent subsidiary. The company has revealed its plan in a note to employees from its CEO Pat Gelsinger, published over a month after Intel disclosed that it's cutting 15 percent of its workforce. Intel is laying off more than 15,000 people as part of its $10 billion cost-reduction plan to regain financial stability following a second-quarter net loss of $1.6 billion. Gelsinger explained in his new memo that turning the foundry into a subsidiary "will unlock important benefits," particularly the ability to evaluate and take external funding directly. 

Gelsinger said that there will be no changes to the foundry's leadership, but the subsidiary will establish its own operating board with independent directors to govern it. According to CNBC, Intel is even considering making the foundry a separate publicly traded company. Intel is in the midst of modernizing its existing fabs and building new ones for its foundry business, which is costing the company billions of dollars, in an effort to catch up to its chipmaking rivals like TSMC and Samsung. The company has reportedly spent around $25 billion a year on its foundry business over the past two years, but that has yet to translate into profit. 

In April, the company revealed in a presentation to investors that the business posted $7 billion in operating losses for 2023, even larger than the $5.2 billion in losses that it incurred the previous year. It had a revenue of $18.9 billion, down 31 percent from its 2022 revenue of $27.49 billion. Gelsinger warned investors at the time that Intel expects its foundry business' operating loss for 2024 to be even bigger and that it doesn't expect to break even until 2027. The foundry's finances aren't the division's only problem: Its next-gen manufacturing process referred to as "18A" reportedly failed crucial tests to prove that it's ready to be used for mass production. 

In addition to announcing that the foundry business will become a subsidiary, Gelsinger also disclosed in the memo that Intel will be selling part of its stake in Altera, another chipmaker that it purchased for $16.7 billion in 2015. 

This article originally appeared on Engadget at https://www.engadget.com/general/intel-is-separating-its-ailing-foundry-business-from-the-main-company-110043046.html?src=rss

Meta bans Russian state media outlets on Facebook and Instagram

Posted on by

Meta has banned RT and other Russian state media outlets on its platforms, which include Facebook and Instagram. "After careful consideration, we expanded our ongoing enforcement against Russian state media outlets," the company told Engadget in a statement. "Rossiya Segodnya, RT and other related entities are now banned from our apps globally for foreign interference activity." Meta shared briefing materials with Reuters, in which the company said it found evidence in the past that Russian outlets took steps to prevent their foreign interference activities online from being detected. It said it expects the outlets to continue with their deceptive practices. The ban will be enforced over the coming days. 

The company's announcement comes days after the Biden administration publicly accused RT of spreading propaganda and disinformation online. Authorities said that RT is spreading propaganda and information to justify and back its invasion of Ukraine using media outlets catering to locals, such as ones targeting Africa and France. 

Meta restricted Russian state media's access to Facebook in Ukraine back in 2022 at the request of its government, in order to limit the spread of those outlets' content. In response, Russia blocked Facebook in the country a few days later and opening a criminal case against the company. A judge later found Facebook and Instagram guilty of "extremist" activity, effectively banning Meta from opening offices and doing business in Russia. 

This article originally appeared on Engadget at https://www.engadget.com/social-media/meta-bans-russian-state-media-outlets-on-facebook-and-instagram-012424410.html?src=rss

23andMe will pay $30 million to settle 2023 data breach lawsuit

Posted on by

23andMe is close to settling a proposed class action lawsuit filed against the company over a data breach that compromised 6.9 million users' information. According to the preliminary settlement filing, the DNA testing company has agreed to pay $30 million to affected customers, as well as to conduct annual computer scans and cybersecurity audits for three years. A website will be built to notify people eligible to a portion of the settlement fund and to facilitate payments. Affected users will also be sent a link where they can delete all their information from the service, and they'll be able to enroll to a three-year Privacy & Medical Shield + Genetic Monitoring program for free. A judge still has to approve those terms. 

In October 2023, the company admitted that the DNA Relatives profile information of roughly 5.5 million customers and the Family Tree profile information of 1.4 million DNA Relative participants had been leaked. It later revealed in a legal filing that the bad actors started breaking into customer accounts in late April 2023 and that they had access to its systems until September that year. It said that the hackers used a technique called credential stuffing, which uses previously compromised login credentials to access customer accounts. 

The breach led to several class action lawsuits filed against the company, including one that accused 23andMe of failing to notify the plaintiffs that they were specifically targeted for having Chinese and Ashkenazi Jewish heritage. In the settlement agreement [PDF] for the consolidated lawsuit, 23andMe noted that it "denies the claims and allegations set forth in the Complaint" and that it "denies that it failed to properly protect the Personal Information of its consumers and users." 

According to Reuters, 23andMe describes its financial condition as "extremely uncertain." In its financial report for the 2024 fiscal year, it revealed that it earned a total revenue of $220 million, down 27 percent from a $299 million revenue the year before. A huge chunk of the settlement money will come from cyber insurance, though, which the company expects to cover $25 million out of the $30 million total. 

This article originally appeared on Engadget at https://www.engadget.com/cybersecurity/23andme-will-pay-30-million-to-settle-2023-data-breach-lawsuit-150058702.html?src=rss

OpenAI is reportedly moving away from its complicated non-profit structure next year

Posted on by

Sam Altman has told OpenAI staff members during their weekly meeting that the company is changing its rather convoluted non-profit corporate structure next year, according to Fortune. The CEO said OpenAI will move away from being controlled by a non-profit entity and will transition into a more traditional for-profit organization. He didn't delve into the specifics of how the company will achieve that goal and what OpenAI's corporate structure will look like exactly. A spokesperson only told Fortune that it remains "focused on building AI that benefits everyone" and that non-profit is "core to [its] mission and will continue to exist."

OpenAI started as a non-profit organization in 2015 that relied on money from donors. In a page explaining its structure, it said that it only raised $130.5 million in total donations over the years, which it says made it clear that "donations alone would not scale with the cost of computational power and talent required to push [its] core research forward." The then-purely non-profit organization created a for-profit subsidiary in order to solve that problem. As Fortune explains, OpenAI's non-profit entity currently controls its for-profit arm, which in turn controls a holding company that takes investments from companies like Microsoft. 

Under this structure, the profit that can be allocated to investors, including Microsoft, has a cap. Anything OpenAI makes beyond the cap will go to its non-profit division. And the company's revenue is booming, according to a report by The Information published in June. OpenAI reportedly doubled its annualized revenue in the first half of the year, thanks to the subscription version of ChatGPT.

The company's complex structure also allowed OpenAI's non-profit board of directors to oust Altman in 2023, because they "no longer [have] confidence in his ability to continue leading OpenAI." Five days later, however, the board was disbanded and replaced, while Altman was reinstated as CEO. 

This article originally appeared on Engadget at https://www.engadget.com/ai/openai-is-reportedly-moving-away-from-its-complicated-non-profit-structure-next-year-130014948.html?src=rss

Meta hides warning labels for AI-edited images

Posted on by

Starting next week, Meta will no longer put an easy-to-see label on Facebook images that were edited using AI tools, and it will make it much harder to determine if they appear in their original state or had been doctored. To be clear, the company will still add a note to AI-edited images, but you'll have to tap on the three-dot menu at the upper right corner of a Facebook post and then scroll down to find "AI Info" among the many other options. Only then will you see the note saying that the content in the post may have been modified with AI. 

Images generated using AI tools, however, will still be marked with an "AI Info" label that can be seen right on the post. Clicking on it will show a note that will say whether it's been labeled because of industry-shared signals or because somebody self-disclosed that it was an AI-generated image. Meta started applying AI-generated content labels to a broader range of videos, audio and images earlier this year. But after widespread complaints from photographers that the company was flagging even non-AI-generated content by mistake, Meta changed the "Made with AI" label wording into "AI Info" by July.

The social network said it worked with companies across the industry to improve its labeling process and that it's making these changes to "better reflect the extent of AI used in content." Still, doctored images are being widely used these days to spread misinformation, and this development could make it trickier to identify false news, which typically pop up more during election season. 

This article originally appeared on Engadget at https://www.engadget.com/social-media/meta-hides-warning-labels-for-ai-edited-images-143004313.html?src=rss

Annapurna’s entire video game team has reportedly left the company

Posted on by

The entire Annapurna Interactive team has left the company after its executives walked out, according to Bloomberg. Apparently, the video game publisher's president, Nathan Gary, had been negotiating with Annapurna Pictures' founder Megan Ellison to spin off Annapurna Interactive into its own entity. A company spokesperson confirmed to Bloomberg that the parties had explored the possibility of a spinoff, but their discussions broke down. Gary and the publisher's other executives had resigned and walked out as a result, and the team's other members had followed suit. 

"All 25 members of the Annapurna Interactive team collectively resigned," the team said in a joint statement. "This was one of the hardest decisions we have ever had to make and we did not take this action lightly."

Annapurna Interactive, like other publishers, teams up with developers, funds their games, takes care of the QA process and then markets and distributes titles when they're ready for the public. Its partner developers have reportedly been scrambling to find out what the team exodus means for them over the past few days. Bloomberg says Hector Sanchez, an Annapurna Interactive co-founder, assured developers that the company will honor their agreements. Sanchez had been at Epic Games the past five years, but he recently rejoined the publisher as its president for Interactive + New Media. Annapurna will also reportedly hire new staff members to replace those who left, essentially building a brand new team. 

"Our top priority is continuing to support our developer and publishing partners during this transition,” Ellison told Bloomberg News. “We’re committed to not only our existing slate of games but also expanding our presence in the interactive space as we continue to look for opportunities to take a more integrated approach to linear and interactive storytelling across film and TV, gaming, and theater."

The publisher made a splash when it debuted with What Remains of Edith Finch back in 2017, and it now has a number of critically acclaimed and popular games under its belt. Its games include the cat simulator Stray, a puzzle adventure game featuring a world-hopping beetle called Cocoon, pop album rhythm action game Sayonara Wild Hearts, and action adventure game Outer Wilds, which is set in a solar system trapped in a time loop. 

This article originally appeared on Engadget at https://www.engadget.com/gaming/annapurnas-entire-video-game-team-has-reportedly-left-the-company-123045395.html?src=rss

Google’s AI notebook can generate a podcast about your notes

Posted on by

Google's latest update for its AI-powered research tool NotebookLM can turn the materials you want to pore over into a podcast-like audio discussion. The new feature called Audio Overview takes information from documents you've uploaded and then generates a "deep dive" discussion between two AI hosts. In addition to summarizing your sources, Google says the hosts will be able to find links between different topics and even banter back and forth. Based on the example the company posted with its announcement, the AI hosts sounded human enough to listen to, though you could still determine that the voices were AI-generated from their inflections and odd pronunciations of certain words. 

Since the feature is still in its experimental stages, Google admits that it has its limitations. The hosts can only speak English at this time, and they sometimes say inaccurate information, which means you will have to double check your material and ensure you didn't just learn something that's not factual. You also can't interrupt the hosts while they're speaking yet, and it still takes several minutes for NotebookLM to generate an Audio Overview for notebooks with larger files. Biao Wang, Google Labs product manager, wrote in the feature's announcement post that his team is "excited to bring audio into NotebookLM" despite those limitations, since they "know some people learn and remember better by listening to conversations."

The company launched NotebookLM back in 2023 as some sort of a digital assistant that you can ask questions about the documents you upload. In June this year, Google announced that NotebookLM has officially started running on Gemini 1.5 Pro, giving it new features and tools, and has expanded to over 200 countries and territories.

This article originally appeared on Engadget at https://www.engadget.com/ai/googles-ai-notebook-can-generate-a-podcast-about-your-notes-140004869.html?src=rss

Polaris Dawn astronauts successfully perform the first commercial spacewalk

Posted on by

Polaris Dawn astronauts Jared Isaacman and Sarah Gillis have successfully exited the confines of their Crew Dragon vehicle to perform the first ever commercial spacewalk on September 12. SpaceX and the Polaris crew started preparing for the event at 6:01AM Eastern time, checking that all their gear was in order and putting on the SpaceX extravehicular activity (EVA) suits that all four of them have to wear, because opening the Dragon's hatch will expose them to outer space. Pure oxygen was pumped into the suits to check for leaks before all the air was let out of the vehicle and before Isaacman opened the hatch. 

A few minutes before 7AM Eastern, Isaacman went through the hatch and exited the vehicle, with the Dragon's camera capturing his silhouette against our planet as the backdrop. He then performed a series of mobility exercises as planned — the whole point of the spacewalk was to test SpaceX's new EVA suits — before going back into the capsule. "Back at home, we all have a lot of work to do," he said as he emerged from the vehicle. "But from here, it looks like a perfect world."

After he went back to his seat, SpaceX engineer Sarah Gillis geared up for her turn. She had to deal with the hatch seal bulging out a bit, but she was ultimately able to climb out of the hatch and do a series of mobility exercises like Isaacman did. Once she was done, the team closed the hatch and started repressurizing Dragon to equalize the atmosphere inside before checking the cabin for any leaks. By 7:55AM, Dragon's repressurization was complete, marking the completion of today's spacewalk. 

This article originally appeared on Engadget at https://www.engadget.com/science/space/polaris-dawn-astronauts-successfully-perform-the-first-commercial-spacewalk-120048616.html?src=rss

How to watch Polaris Dawn astronauts attempt the first commercial spacewalk

Posted on by

Two of Polaris Dawn's four astronauts could make history today by performing the first ever commercial spacewalk at around 700 kilometers (435 miles) above our planet. Jared Isaacman, the mission leader and funder, and SpaceX engineer Sarah Gillis will leave the confines and safety of the Crew Dragon capsule for 15 and 20 minutes each. The other two crew members, Scott Poteet and Anna Menon, will stay inside the vehicle to monitor the spacewalkers' status and make sure everything's in order. SpaceX will stream the event live on its website and on X starting at 4:50AM Eastern time. In case the spacewalk has to be called off, the astronauts will have another opportunity on September 13 at the same time. 

As The New York Times explains, the Crew Dragon has no airlock like the International Space Station does, so the astronauts will have to let the all the air out of the vehicle before opening one of its hatches. All four will have to wear the company's extravehicular activity (EVA) suits, which are upgraded and tougher versions of its suits for intravehicular activity (IVA). 

SpaceX's EVAs come with new joints that can flex and rotate, thereby providing greater mobility. Their helmets are equipped with a heads-up display (HUD) and camera, and the spacesuits as a whole have a Faraday layer that can protect them from electric fields. The Polaris Dawn crew's spacewalk will put the suit to the test as they subject it to the harsh environment of outer space while conducting mobility tests. "The development of this suit and the execution of the spacewalk will be important steps toward a scalable design for spacesuits on future long-duration missions as life becomes multiplanetary," the Polaris Dawn's website reads. The whole operation, from the beginning until the astronauts close the hatch to repressurize Crew Dragon, will last for two hours.

The civilian mission launched in the early hours of September 10 after multiple delays. In addition to accomplishing the first commercial spacewalk, the mission has other objectives, including sending a crew farther than any Dragon mission before and farther than anybody has been since the Apollo program, even reaching parts of the Van Allen radiation belt. 

This article originally appeared on Engadget at https://www.engadget.com/science/space/how-to-watch-polaris-dawn-astronauts-attempt-the-first-commercial-spacewalk-074548007.html?src=rss

Meta scraped every Australian user’s account to train its AI

Posted on by

In a government inquiry about AI adoption in Australia, Meta's global privacy director Melinda Claybaugh was asked whether her company has been collecting Australians' data to train its generative AI technology. According to ABC News, Claybaugh initially denied the claim, but upon being pressed, she ultimately admitted that Meta scrapes all the photos and texts in all Facebook and Instagram posts from as far back as 2007, unless the user had set their posts to private. Further, she admitted that the company isn't offering Australians an opt-out option like it does to users in the European Union. 

Claybaugh said that Meta doesn't scrape the accounts of users under 18 years old, but she admitted that the company still collects their photos and other information if they're posted on their parents' or guardians' accounts. She couldn't answer, however, if the company collects data from previous years once a user turns 18. Upon being asked why Meta doesn't offer Australians the option not to consent to data collection, Claybaugh said that it exists in the EU "in response to a very specific legal frame," which most likely pertains to the bloc's General Data Protection Regulation (GDPR).

Meta had notified users in the EU that it will collect their data for AI training unless they opt out. "I will say that the ongoing conversation in Europe is the direct result of the existing regulatory landscape," Claybaugh explained during the inquiry. But even in the region, Claybaugh said that there's an "ongoing legal question around what is the interpretation of existing privacy law with respect to AI training." Meta decided not to offer its multimodal AI model and future versions in the block due to what it says is a lack of clarity from European regulators. Most of its concerns centered around the difficulties of training AI models with data from European users while complying with GDPR rules. 

Despite those legal questions around AI adoption in Europe, bottom line is that Meta is giving users in the bloc the power to block data collection. "Meta made it clear today that if Australia had these same laws Australians' data would also have been protected," Australian Senator David Shoebridge told ABC News. "The government's failure to act on privacy means companies like Meta are continuing to monetise and exploit pictures and videos of children on Facebook."

This article originally appeared on Engadget at https://www.engadget.com/apps/meta-scraped-every-australian-users-account-to-train-its-ai-120026200.html?src=rss