Author name: Shannon Garcia

youtube-will-require-disclosure-of-ai-manipulated-videos-from-creators

YouTube will require disclosure of AI-manipulated videos from creators

You could also just ban manipulations altogether? —

YouTube wants “realistic” likenesses or audio fabrications to be labeled.

YouTube will require disclosure of AI-manipulated videos from creators

Future Publishing | Getty Images

YouTube is rolling out a new requirement for content creators: You must disclose when you’re using AI-generated content in your videos. The disclosure appears in the video upload UI and will be used to power an “altered content” warning on videos.

Google previewed the “misleading AI content” policy in November, but the questionnaire is now going live. Google is mostly concerned about altered depictions of real people or events, which sounds like more election-season concerns about how AI can mislead people. Just last week, Google disabled election questions for its “Gemini” chatbot.

As always, the exact rules on YouTube are up for interpretation. Google says it’s “requiring creators to disclose to viewers when realistic content—content a viewer could easily mistake for a real person, place, or event—is made with altered or synthetic media, including generative AI,” but doesn’t require creators to disclose manipulated content that is “clearly unrealistic, animated, includes special effects, or has used generative AI for production assistance.”

Google gives examples of when a disclosure is necessary, and the new video upload questionnaire walks content creators through these requirements:

  • Using the likeness of a realistic person: Digitally altering content to replace the face of one individual with another’s or synthetically generating a person’s voice to narrate a video.
  • Altering footage of real events or places: Such as making it appear as if a real building caught fire, or altering a real cityscape to make it appear different from reality.
  • Generating realistic scenes: Showing a realistic depiction of fictional major events, like a tornado moving toward a real town.
  • Google’s video upload questionnaire.

    YouTube

  • Take note of the super-tiny message at the bottom, denoting “altered or synthetic content.”

    YouTube

  • You can expand the description for slightly more info.

    YouTube

Google says the labels will start rolling out “across all YouTube surfaces and formats in the weeks ahead, beginning with the YouTube app on your phone, and soon on your desktop and TV.” The company says it’s also working on a process for people who are the subject of an AI-manipulated video to request its removal, but it doesn’t have details on that yet.

YouTube will require disclosure of AI-manipulated videos from creators Read More »

deepmind-co-founder-mustafa-suleyman-will-run-microsoft’s-new-consumer-ai-unit

DeepMind co-founder Mustafa Suleyman will run Microsoft’s new consumer AI unit

Minding deeply —

Most staffers from Suleyman’s startup, Inflection, will join Microsoft as well.

Mustafa Suleyman, talks on Day 1 of the AI Safety Summit at Bletchley Park at Bletchley Park on November 1, 2023 in Bletchley, England.

Enlarge / Mustafa Suleyman, talks on Day 1 of the AI Safety Summit at Bletchley Park at Bletchley Park on November 1, 2023 in Bletchley, England.

Microsoft has hired Mustafa Suleyman, the co-founder of Google’s DeepMind and chief executive of artificial intelligence start-up Inflection, to run a new consumer AI unit.

Suleyman, a British entrepreneur who co-founded DeepMind in London in 2010, will report to Microsoft chief executive Satya Nadella, the company announced on Tuesday. He will launch a division of Microsoft that brings consumer-facing products including Microsoft’s Copilot, Bing, Edge, and GenAI under one team called Microsoft AI.

It is the latest move by Microsoft to capitalize on the boom in generative AI. It has invested $13 billion in OpenAI, the maker of ChatGPT, and rapidly integrated its technology into Microsoft products.

Microsoft’s investment in OpenAI has given it an early lead in Silicon Valley’s race to deploy AI, leaving its biggest rival, Google, struggling to catch up. It also has invested in other AI startups, including French developer Mistral.

It has been rolling out an AI assistant in its products such as Windows, Office software, and cyber security tools. Suleyman’s unit will work on projects including integrating an AI version of Copilot into its Windows operating system and enhancing the use of generative AI in its Bing search engine.

Nadella said in a statement on Tuesday: “I’ve known Mustafa for several years and have greatly admired him as a founder of both DeepMind and Inflection, and as a visionary, product maker and builder of pioneering teams that go after bold missions.”

DeepMind was acquired by Google in 2014 for $500 million, one of the first large bets by a big tech company on a startup AI lab. The company faced controversy a few years later over some of its projects, including its work for the UK healthcare sector, which was found by a government watchdog to have been granted inappropriate access to patient records.

Suleyman, who was the main public face for the company, was placed on leave in 2019. DeepMind workers had complained that he had an overly aggressive management style. Addressing staff complaints at the time, Suleyman said: “I really screwed up. I was very demanding and pretty relentless.”

He moved to Google months later, where he led AI product management. In 2022, he joined Silicon Valley venture capital firm Greylock and launched Inflection later that year.

Microsoft will also hire most of Inflection’s staff, including Karén Simonyan, cofounder and chief scientist of Inflection, who will be chief scientist of the AI group. Microsoft did not clarify the number of employees moving over but said it included AI engineers, researchers, and large language model builders who have designed and co-authored “many of the most important contributions in advancing AI over the last five years.”

Inflection, a rival to OpenAI, will switch its focus from its consumer chatbot, Pi, and instead move to sell enterprise AI software to businesses, according to a statement on its website. Sean White, who has held various technology roles, has joined as its new chief executive.

Inflection’s third cofounder, Reid Hoffman, the founder and executive chair of LinkedIn, will remain on Inflection’s board. Inflection had raised $1.3 billion in June, valuing the group at about $4 billion, in one of the largest fundraisings by an AI start-up amid an explosion of interest in the sector.

The new unit marks a big organizational shift at Microsoft. Mikhail Parakhin, its president of web services, will move along with his entire team to report to Suleyman.

“We have a real shot to build technology that was once thought impossible and that lives up to our mission to ensure the benefits of AI reach every person and organization on the planet, safely and responsibly,” Nadella said.

Competition regulators in the US and Europe have been scrutinising the relationship between Microsoft and OpenAI amid a broader inquiry into AI investments.

© 2024 The Financial Times Ltd. All rights reserved. Not to be redistributed, copied, or modified in any way.

DeepMind co-founder Mustafa Suleyman will run Microsoft’s new consumer AI unit Read More »

office-2024-will-be-the-next-standalone-release,-as-the-office-brand-lives-on

Office 2024 will be the next standalone release, as the Office brand lives on

return to office —

Consumer prices stay the same; MS promises at least one more standalone release.

Office 2024 will be the next standalone release, as the Office brand lives on

Microsoft

Last week, Microsoft announced that it would soon begin offering previews of Microsoft Office 2024, the next standalone perpetually licensed version of the Office suite. Like Office 2021 before it, Office 2024 will be part of Microsoft’s Long-Term Servicing Channel (LTSC), which is intended for IT administrators and users who value stability and predictability over constant iteration.

But Microsoft is being clearer than ever that it would really like people to move to using Microsoft 365 subscriptions, referring to Office 2024 as “a specialty product that Microsoft has committed to maintaining for use in exceptional circumstances.” The company will be increasing prices for businesses by “up to 10 percent” compared to Office 2021, a price hike that Microsoft says will “support continued innovation in this niche space.” Pricing for the consumer version of Office 2024 should stay the same as it is for Office 2021.

Office 2024 will receive support and security updates for five years from its release date, which will be “later this year,” along with a new LTSC release of Windows 11. The company has also committed to releasing at least one more standalone version of Office in the future. If you bought Office 2021 and you’re still happy with it, you’ll still get support (including security updates) until October of 2026. Support for Office 2019 ended in October 2023.

Highlights of the Office 2024 release include “new meeting creation options and search enhancements in Outlook, dozens of new Excel features and functions including Dynamic Charts and Arrays; and improved performance, security, and accessibility,” according to Microsoft. One missing feature of note will be Microsoft Publisher, which will be discontinued in October 2026.

Like other standalone Office releases, Office 2024’s feature set will be frozen in time without the continuous changes present in the Microsoft 365 versions of the same apps. And while the perpetually licensed versions of Office can interact with Microsoft services like OneDrive, the company says that other features like the Copilot AI assistant won’t be available in Office 2024 because it is a “disconnected product.” It’s also missing real-time collaboration features available in Microsoft 365.

Is it a rebrand if you keep using the old brand name in parentheses?

Enlarge / Is it a rebrand if you keep using the old brand name in parentheses?

Microsoft

Among other things, the continued existence of the standalone Office product shows that the “Office” brand name will still be with us for a while, despite Microsoft’s formal decision to retire it back in late 2022. “Office” remains useful as a noun that refers to all of these apps collectively—and separately from the Microsoft 365 subscription product. Even in places where the Office name has been removed, Microsoft seems reticent to do away with it entirely; the Microsoft 365 app that comes built into Windows has the new name and the new logo but also includes “Office” in parentheses as if to say “yes, we know what name most people use to talk about this software.”

Office 2024 will be the next standalone release, as the Office brand lives on Read More »

f1’s-pursuit-of-sustainability-drives-pirelli-to-unveil-forest-friendly-tires

F1’s pursuit of sustainability drives Pirelli to unveil forest-friendly tires

pit stop —

The Forest Stewardship Council has given its approval to Pirelli’s natural rubber.

A pirelli F1 tire with the FSC logo on it

Enlarge / You’ll notice the Forest Stewardship Council’s logo on the sidewall to the right of the Pirelli logo.

Pirelli

Formula 1 is on a big sustainability kick. The race cars are switching to carbon neutral synthetic fuels. Teams are improving their logistics to cut freight emissions. Race tracks are starting to run entirely on solar power. And now, the tires that Pirelli brings to the races have been given the seal of approval by an NGO as meeting its standards for sustainable forestry.

It will be hard to spot when the cars are moving, but this year, you’ll find a tree logo on the sidewall. That indicates that the natural rubber that went into making the tire has been certified by the Forest Stewardship Council. Natural rubber makes up about 15 percent of the rubber in an F1 tire, with the rest being synthetic.

According to the FSC, natural rubber is a key driver of deforestry, as well as human rights abuses, particularly among the smallholders who farm 85 percent of the world’s natural rubber. By putting its logo on the tire, the FSC says that Pirelli has met “the world’s most credible standards for sustainable forestry,” protecting both the forests and the forest communities’ rights, including fair wages.

It’s one of a number of steps that Pirelli has put in place to make its F1 program more sustainable.

“I believe that the certification is an important step in this direction because it’s not Pirelli that is certifiying itself; it is a recognized third party that is giving us this certification, from the way in which we collect natural rubber, with respect of biodiversity, respect of the local population, the way we transport or use the natural rubber,” explained Mario Isola, head of Pirelli’s F1 program.

The synthetic rubber—chosen because it allows Pirelli to tune the characteristics it needs for the tires’ performance—is another area of attention. “Our R&D is focused on replacing the current material with more sustainable materials, keeping the same level of performance characteristics of the tire,” Isola told Ars.

Pirelli technicians work on the tires during practice ahead of the F1 Grand Prix of Saudi Arabia at Jeddah Corniche Circuit on March 7, 2024, in Jeddah, Saudi Arabia.

Enlarge / Pirelli technicians work on the tires during practice ahead of the F1 Grand Prix of Saudi Arabia at Jeddah Corniche Circuit on March 7, 2024, in Jeddah, Saudi Arabia.

Qian Jun/MB Media/Getty Images

In other racing disciplines, particularly sports car racing, series have begun to restrict the total tire allocation across a race weekend to drive the development of more durable tires that will be used across multiple stints rather than being changed at each pit stop. That’s less appropriate in F1, where the rules require using two different tire compounds during a race. But for this year, Pirelli may well be able to cut the number of wet and intermediate tires by half.

“We are working on another idea that is what we call ‘strip and fit.’ When we fit a tire on a rim, even if it is new, we have to scrap it because of the bead and the stress that you put on the bead,” Isola said.

“But we made an investigation on wet and intermediate tires where the stress on the tire was lower compared to the slick tire. So the tires that we are going to fit but not use during the first half of the season will be dismounted and checked, and then we can use them in the second half of the season. If it doesn’t rain—obviously, we cannot control the weather—we are going to save roughly 50 percent of the rain tires,” he told me.

In other F1 tire news, we’ve now learned that the sport will stick with 18-inch wheels when the technical regulations undergo their next shake-up ahead of the 2026 season.

F1 only moved to 18-inch wheels from much smaller 13-inch wheels at the start of the 2022 season, long after any new vehicle was equipped with wheels so small. There have been complaints that the larger 18-inch wheels have added too much unsprung weight to the current generation of F1 cars, which are by far the heaviest the sport has seen in its history.

Consequently, it was believed that the sport might reduce the wheel size to 16 inches in 2026. But that would require an expensive testing program, and since 16-inch wheels are barely more road-relevant to current new vehicles than 13-inch wheels, the decision was made to stick with what we mostly have now, although the final tire size and shape have yet to be decided upon.

F1’s pursuit of sustainability drives Pirelli to unveil forest-friendly tires Read More »

facebook,-instagram-may-cut-fees-by-nearly-50%-in-scramble-for-dma-compliance

Facebook, Instagram may cut fees by nearly 50% in scramble for DMA compliance

Facebook, Instagram may cut fees by nearly 50% in scramble for DMA compliance

Meta is considering cutting monthly subscription fees for Facebook and Instagram users in the European Union nearly in half to comply with the Digital Market Act (DMA), Reuters reported.

During a day-long public workshop on Meta’s DMA compliance, Meta’s competition and regulatory director, Tim Lamb, told the European Commission (EC) that individual subscriber fees could be slashed from 9.99 euros to 5.99 euros. Meta is hoping that reducing fees will help to speed up the EC’s process for resolving Meta’s compliance issues. If Meta’s offer is accepted, any additional accounts would then cost 4 euros instead of 6 euros.

Lamb said that these prices are “by far the lowest end of the range that any reasonable person should be paying for services of these quality,” calling it a “serious offer.”

The DMA requires that Meta’s users of Facebook, Instagram, Facebook Messenger, and Facebook Marketplace “freely” give consent to share data used for ad targeting without losing access to the platform if they’d prefer not to share data. That means services must provide an acceptable alternative for users who don’t consent to data sharing.

“Gatekeepers should enable end users to freely choose to opt-in to such data processing and sign-in practices by offering a less personalized but equivalent alternative, and without making the use of the core platform service or certain functionalities thereof conditional upon the end user’s consent,” the DMA says.

Designated gatekeepers like Meta have debated what it means for a user to “freely” give consent, suggesting that offering a paid subscription for users who decline to share data would be one route for Meta to continue offering high-quality services without routinely hoovering up data on all its users.

But EU privacy advocates like NOYB have protested Meta’s plan to offer a subscription model instead of consenting to data sharing, calling it a “pay or OK model” that forces Meta users who cannot pay the fee to consent to invasive data sharing they would otherwise decline. In a statement shared with Ars, NOYB chair Max Schrems said that even if Meta reduced its fees to 1.99 euros, it would be forcing consent from 99.9 percent of users.

“We know from all research that even a fee of just 1.99 euros or less leads to a shift in consent from 3–10 percent that genuinely want advertisement to 99.9 percent that still click yes,” Schrems said.

In the EU, the General Data Protection Regulation (GDPR) “requires that consent must be ‘freely’ given,” Schrems said. “In reality, it is not about the amount of money—it is about the ‘pay or OK’ approach as a whole. The entire purpose of ‘pay or OK’, is to get users to click on OK, even if this is not their free and genuine choice. We do not think the mere change of the amount makes this approach legal.”

Where EU stands on subscription models

Meta expects that a subscription model is a legal alternative under the DMA. The tech giant said it was launching EU subscriptions last November after the Court of Justice of the European Union (CJEU) “endorsed the subscriptions model as a way for people to consent to data processing for personalized advertising.”

It’s unclear how popular the subscriptions have been at the current higher cost. Right now in the EU, monthly Facebook and Instagram subscriptions cost 9.99 euros per month on the web or 12.99 euros per month on iOS and Android, with additional fees of 6 euros per month on the web and 8 euros per month on iOS and Android for each additional account. Meta declined to comment on how many EU users have subscribed, noting to Ars that it has no obligation to do so.

In the CJEU case, the court was reviewing Meta’s GDPR compliance, which Schrems noted is less strict than the DMA. The CJEU specifically said that under the GDPR, “users must be free to refuse individually”—”in the context of” signing up for services— “to give their consent to particular data processing operations not necessary” for Meta to provide such services “without being obliged to refrain entirely from using the service.”

Facebook, Instagram may cut fees by nearly 50% in scramble for DMA compliance Read More »

health-experts-plead-for-unvaxxed-americans-to-get-measles-shot-as-cases-rise

Health experts plead for unvaxxed Americans to get measles shot as cases rise

MMR is safe and effective —

The US hit last year’s total in under 12 weeks, suggesting we’re in for a bad time.

A view from a hospital as children receiving medical treatment, in capital Kabul, Afghanistan on April 18, 2022. More than 130 children have died from the measles in Afghanistan since the beginning of this year.

Enlarge / A view from a hospital as children receiving medical treatment, in capital Kabul, Afghanistan on April 18, 2022. More than 130 children have died from the measles in Afghanistan since the beginning of this year.

The Centers for Disease Control and Prevention and the American Medical Association sent out separate but similar pleas on Monday for unvaccinated Americans to get vaccinated against the extremely contagious measles virus as vaccination rates have slipped, cases are rising globally and nationally, and the spring-break travel period is beginning.

In the first 12 weeks of 2024, US measles cases have already matched and likely exceeded the case total for all of 2023. According to the CDC, there were 58 measles cases reported from 17 states as of March 14. But media tallies indicate there have been more cases since then, with at least 60 cases now in total, according to CBS News. In 2023, there were 58 cases in 20 states.

“As evident from the confirmed measles cases reported in 17 states so far this year, when individuals are not immunized as a matter of personal preference or misinformation, they put themselves and others at risk of disease—including children too young to be vaccinated, cancer patients, and other immunocompromised people,” AMA President Jesse Ehrenfeld said in a statement urging vaccination Monday.

The latest data indicates that vaccination rates among US kindergarteners have slipped to 93 percent nationally, below the 95 percent target to prevent the spread of the disease. And vaccine exemptions for non-medical reasons have reached an all-time high.

The CDC released a health advisory on Monday also urging measles vaccination. The CDC drove home the point that unvaccinated Americans are largely responsible for importing the virus, and pockets of unvaccinated children in local communities spread it once it’s here. The 58 measles infections that have been reported to the agency so far include cases from seven outbreaks in seven states. Most of the cases are in vaccine-eligible children aged 12 months and older who are unvaccinated. Of the 58 cases, 54 (93 percent) are linked to international travel, and most measles importations are by unvaccinated US residents who travel abroad and bring measles home with them, the CDC flagged.

The situation is likely to worsen as Americans begin spring travel, the CDC suggested. “Many countries, including travel destinations such as Austria, the Philippines, Romania, and the United Kingdom, are experiencing measles outbreaks,” the CDC said. “To prevent measles infection and reduce the risk of community transmission from importation, all US residents traveling internationally, regardless of destination, should be current on their [measles-mumps-rubella (MMR)] vaccinations.” The agency added in a recommendation to parents that “even if not traveling, ensure that children receive all recommended doses of MMR vaccine. Two doses of MMR vaccine provide better protection (97 percent) against measles than one dose (93 percent). Getting MMR vaccine is much safer than getting measles, mumps, or rubella.”

For Americans who are already vaccinated and communities with high vaccination coverage, the risk is low, the CDC noted. “However, pockets of low coverage leave some communities at higher risk for outbreaks.” This, in turn, threatens wider, continuous spread that could overturn the country’s status of having eliminated measles, which was declared in 2000. The US was close to losing its elimination status in 2019 when outbreaks among unvaccinated children drove 1,247 cases across 31 states. Vaccination rates have only fallen since then.

“The reduction in measles vaccination threatens to erase many years of progress as this previously eliminated vaccine-preventable disease returns,” the AMA’s Ehrenfeld warned.

As Ars has reported previously, measles is among the most contagious viruses known and can linger in airspace for up to two hours. Up to 90 percent of unvaccinated people exposed will contract it. Symptoms can include high fever, runny nose, red and watery eyes, and a cough, as well as the hallmark rash. About 1 in 5 unvaccinated people with measles are hospitalized, while 1 in 20 infected children develop pneumonia, and up to 3 in 1,000 children die of the infection. Brain swelling (encephalitis) can occur in 1 in 1,000 children, which can lead to hearing loss and intellectual disabilities. The virus can also destroy immune responses to previous infections—a phenomenon known as “immune amnesia”—which can leave children vulnerable to various other infections for years afterward.

Health experts plead for unvaxxed Americans to get measles shot as cases rise Read More »

nvidia-unveils-blackwell-b200,-the-“world’s-most-powerful-chip”-designed-for-ai

Nvidia unveils Blackwell B200, the “world’s most powerful chip” designed for AI

There’s no knowing where we’re rowing —

208B transistor chip can reportedly reduce AI cost and energy consumption by up to 25x.

The GB200

Enlarge / The GB200 “superchip” covered with a fanciful blue explosion.

Nvidia / Benj Edwards

On Monday, Nvidia unveiled the Blackwell B200 tensor core chip—the company’s most powerful single-chip GPU, with 208 billion transistors—which Nvidia claims can reduce AI inference operating costs (such as running ChatGPT) and energy consumption by up to 25 times compared to the H100. The company also unveiled the GB200, a “superchip” that combines two B200 chips and a Grace CPU for even more performance.

The news came as part of Nvidia’s annual GTC conference, which is taking place this week at the San Jose Convention Center. Nvidia CEO Jensen Huang delivered the keynote Monday afternoon. “We need bigger GPUs,” Huang said during his keynote. The Blackwell platform will allow the training of trillion-parameter AI models that will make today’s generative AI models look rudimentary in comparison, he said. For reference, OpenAI’s GPT-3, launched in 2020, included 175 billion parameters. Parameter count is a rough indicator of AI model complexity.

Nvidia named the Blackwell architecture after David Harold Blackwell, a mathematician who specialized in game theory and statistics and was the first Black scholar inducted into the National Academy of Sciences. The platform introduces six technologies for accelerated computing, including a second-generation Transformer Engine, fifth-generation NVLink, RAS Engine, secure AI capabilities, and a decompression engine for accelerated database queries.

Press photo of the Grace Blackwell GB200 chip, which combines two B200 GPUs with a Grace CPU into one chip.

Enlarge / Press photo of the Grace Blackwell GB200 chip, which combines two B200 GPUs with a Grace CPU into one chip.

Several major organizations, such as Amazon Web Services, Dell Technologies, Google, Meta, Microsoft, OpenAI, Oracle, Tesla, and xAI, are expected to adopt the Blackwell platform, and Nvidia’s press release is replete with canned quotes from tech CEOs (key Nvidia customers) like Mark Zuckerberg and Sam Altman praising the platform.

GPUs, once only designed for gaming acceleration, are especially well suited for AI tasks because their massively parallel architecture accelerates the immense number of matrix multiplication tasks necessary to run today’s neural networks. With the dawn of new deep learning architectures in the 2010s, Nvidia found itself in an ideal position to capitalize on the AI revolution and began designing specialized GPUs just for the task of accelerating AI models.

Nvidia’s data center focus has made the company wildly rich and valuable, and these new chips continue the trend. Nvidia’s gaming GPU revenue ($2.9 billion in the last quarter) is dwarfed in comparison to data center revenue (at $18.4 billion), and that shows no signs of stopping.

A beast within a beast

Press photo of the Nvidia GB200 NVL72 data center computer system.

Enlarge / Press photo of the Nvidia GB200 NVL72 data center computer system.

The aforementioned Grace Blackwell GB200 chip arrives as a key part of the new NVIDIA GB200 NVL72, a multi-node, liquid-cooled data center computer system designed specifically for AI training and inference tasks. It combines 36 GB200s (that’s 72 B200 GPUs and 36 Grace CPUs total), interconnected by fifth-generation NVLink, which links chips together to multiply performance.

A specification chart for the Nvidia GB200 NVL72 system.

Enlarge / A specification chart for the Nvidia GB200 NVL72 system.

“The GB200 NVL72 provides up to a 30x performance increase compared to the same number of NVIDIA H100 Tensor Core GPUs for LLM inference workloads and reduces cost and energy consumption by up to 25x,” Nvidia said.

That kind of speed-up could potentially save money and time while running today’s AI models, but it will also allow for more complex AI models to be built. Generative AI models—like the kind that power Google Gemini and AI image generators—are famously computationally hungry. Shortages of compute power have widely been cited as holding back progress and research in the AI field, and the search for more compute has led to figures like OpenAI CEO Sam Altman trying to broker deals to create new chip foundries.

While Nvidia’s claims about the Blackwell platform’s capabilities are significant, it’s worth noting that its real-world performance and adoption of the technology remain to be seen as organizations begin to implement and utilize the platform themselves. Competitors like Intel and AMD are also looking to grab a piece of Nvidia’s AI pie.

Nvidia says that Blackwell-based products will be available from various partners starting later this year.

Nvidia unveils Blackwell B200, the “world’s most powerful chip” designed for AI Read More »

centurylink-left-customers-without-internet-for-39-days—until-ars-stepped-in

CenturyLink left customers without Internet for 39 days—until Ars stepped in

A

Aurich Lawson | Getty Images

When a severe winter storm hit Oregon on January 13, Nicholas Brown’s CenturyLink fiber Internet service stopped working at his house in Portland.

The initial outage was understandable amid the widespread damage caused by the storm, but CenturyLink’s response was poor. It took about 39 days for CenturyLink to restore broadband service to Brown and even longer to restore service to one of his neighbors. Those reconnections only happened after Ars Technica contacted the telco firm on the customers’ behalf last week.

Brown had never experienced any lengthy outage in over four years of subscribing to CenturyLink, so he figured the telco firm would restore his broadband connection within a reasonable amount of time. “It had practically never gone down at all up to this point. I’ve been quite happy with it,” he said.

While CenturyLink sent trucks to his street to reconnect most of his neighbors after the storm and Brown regularly contacted CenturyLink to plead for a fix, his Internet connection remained offline. Brown had also lost power, but the electricity service was reconnected within about 48 hours, while the broadband service remained offline for well over a month.

Fearing he had exhausted his options, Brown contacted Ars. We sent an email to CenturyLink’s media department on February 21 to seek information on why the outage lasted so long.

Telco finally springs into action

Roughly four hours after we contacted the firm, a CenturyLink technician arrived at the Portland house Brown shares with his partner, Jolene Edwards. The technician was able to reconnect them that day.

“At 4: 30 pm, a CenturyLink tech showed up unannounced,” Brown told us. “No one was home at the time, but he said he would wait. I get the idea that he was told not to come back until it was fixed.”

Brown’s neighbor, Leonard Bentz, also lost Internet access on January 13 and remained offline for two days longer than Brown. The technician who arrived on February 21 didn’t reconnect Bentz’s house.

“My partner gently tried to egg him to go over there and fix them too, and he more or less said, ‘That’s not the ticket that I have,'” Brown said.

After getting Bentz’s name and address, we contacted CenturyLink again on February 22 to notify them that he also needed to be reconnected. CenturyLink later confirmed to us that it restored his Internet service on February 23.

“They kept putting me off and putting me off”

Bentz told Ars that during the month-plus outage, he called CenturyLink several times. Customer service reps and a supervisor told him the company would send someone to fix his service, but “they kept putting me off and putting me off and putting me off,” Bentz said.

On one of those calls, Bentz said that CenturyLink promised him seven free months of service in exchange for the long outage. Brown told us he received a refund for the entire length of his outage, plus a bit extra. He pays $65 a month for gigabit service.

Brown said he is “happy enough with the resolution,” at least financially since he “got all the money for the non-service.” But those 39 days without Internet service will remain a bad memory.

Unfortunately, Internet service providers like CenturyLink have a history of failing to fix problems until media coverage exposes their poor customer service. CenturyLink is officially called Lumen these days, but it still uses the CenturyLink brand name.

After fixing Brown’s service in Portland, a CenturyLink spokesperson gave us the following statement:

It’s frustrating to have your services down and for that we apologize. We’ve brought in additional resources to assist in restoring service that was knocked out due to severe storms and multiple cases of vandalism. Some services are back, and we are working diligently to completely restore everything. In fact, we have technicians there now. We appreciate our customers’ patience and understanding, and we welcome calls from our customers to discuss their service.

CenturyLink left customers without Internet for 39 days—until Ars stepped in Read More »

a-big-boost-to-europe’s-climate-change-goals

A big boost to Europe’s climate-change goals

carbon-neutral continent —

A new policy called CBAM will assist Europe’s ambition to become carbon-neutral.

Steelworker starting molten steel pour in steelworks facility.

Enlarge / Materials such as steel, cement, aluminum, electricity, fertilizer, hydrogen, and iron will soon be subject to greenhouse gas emissions fees when imported into Europe.

Monty Rakusen/Getty

The year 2023 was a big one for climate news, from record heat to world leaders finally calling for a transition away from fossil fuels. In a lesser-known milestone, it was also the year the European Union soft-launched an ambitious new initiative that could supercharge its climate policies.

Wrapped in arcane language studded with many a “thereof,” “whereas” and “having regard to” is a policy that could not only help fund the European Union’s pledge to become the world’s first carbon-neutral continent, but also push industries all over the world to cut their carbon emissions.

It’s the establishment of a carbon price that will force many heavy industries to pay for each ton of carbon dioxide, or equivalent emissions of other greenhouse gases, that they emit. But what makes this fee revolutionary is that it will apply to emissions that don’t happen on European soil. The EU already puts a price on many of the emissions created by European firms; now, through the new Carbon Border Adjustment Mechanism, or CBAM, the bloc will charge companies that import the targeted products—cement, aluminum, electricity, fertilizer, hydrogen, iron, and steel—into the EU, no matter where in the world those products are made.

These industries are often large and stubborn sources of greenhouse gas emissions, and addressing them is key in the fight against climate change, says Aaron Cosbey, an economist at the International Institute for Sustainable Development, an environmental think tank. If those companies want to continue doing business with European firms, they’ll have to clean up or pay a fee. That creates an incentive for companies worldwide to reduce emissions.

In CBAM’s first phase, which started in October 2023, companies importing those materials into the EU must report on the greenhouse gas emissions involved in making the products. Beginning in 2026, they’ll have to pay a tariff.

Even having to supply emissions data will be a big step for some producers and could provide valuable data for climate researchers and policymakers, says Cosbey.

“I don’t know how many times I’ve gone through this exercise of trying to identify, at a product level, the greenhouse gas intensity of exports from particular countries and had to go through the most amazing, torturous processes to try to do those estimates,” he says. “And now it’s going to be served to me on a plate.”

CBAM will apply to a set of products that are linked to heavy greenhouse gas emissions.

Enlarge / CBAM will apply to a set of products that are linked to heavy greenhouse gas emissions.

Side benefits at home

While this new carbon price targets companies abroad, it will also help the EU to pursue its climate ambitions at home. For one thing, the extra revenues could go toward financing climate-friendly projects and promising new technologies.

But it also allows the EU to tighten up on domestic pollution. Since 2005, the EU has set a maximum, or cap, on the emissions created by a range of industrial “installations” such as oil and metal refineries. It makes companies within the bloc use credits, or allowances, for each ton of carbon dioxide—or equivalent discharges of other greenhouse gases—that they emit, up to that cap. Some allowances are currently granted for free, but others are bought at auction or traded with other companies in a system known as a carbon market.

But this idea—of making it expensive to harm the planet—creates a conundrum. If doing business in Europe becomes too expensive, European industry could flee the continent for countries that don’t have such high fees or strict regulations. That would damage the European economy and do nothing to solve the environmental crisis. The greenhouse gases would still be emitted—perhaps more than if the products had been made in Europe—and climate change would careen forward on its destructive path.

The Carbon Border Adjustment Mechanism aims to impose the same carbon price for products made abroad as domestic producers must pay under the EU’s system. In theory, that keeps European businesses competitive with imports from international rivals. It also addresses environmental concerns by nudging companies overseas toward reducing greenhouse gas emissions rather than carrying on as usual.

This means the EU can further tighten up its carbon market system at home. With international competition hopefully less of a concern, it plans to phase out some leniencies, such as some of the free emission allowances, that existed to help keep domestic industries competitive.

That’s a big deal, says Cosbey. Dozens of countries have carbon pricing systems, but they all create exceptions to keep heavy industry from getting obliterated by international competition. The carbon border tariff could allow the EU to truly force its industries—and consumers—to pay the price, he says.

“That is ambitious; nobody in the world is doing that.”

A big boost to Europe’s climate-change goals Read More »

$30-doorbell-cameras-have-multiple-serious-security-flaws,-says-consumer-reports

$30 doorbell cameras have multiple serious security flaws, says Consumer Reports

Video doorbell security —

Models still widely available on e-commerce sites after issues reported.

Image showing a delivery person saying

Enlarge / Consumer Reports’ investigation suggests that, should this delivery person press and hold the bell button and then pair using Eken’s app, he could see if other delivery people get such a perfunctory response.

Eken

Video doorbell cameras have been commoditized to the point where they’re available for $30–$40 on marketplaces like Amazon, Walmart, Temu, and Shein. The true cost of owning one might be much greater, however.

Consumer Reports (CR) has released the findings of a security investigation into two budget-minded doorbell brands, Eken and Tuck, which are largely the same hardware produced by the Eken Group in China, according to CR. The cameras are further resold under at least 10 more brands. The cameras are set up through a common mobile app, Aiwit. And the cameras share something else, CR claims: “troubling security vulnerabilities.”

The pairing procedure for one of Eken's doorbell cameras, which allows a malicious actor quite a bit of leeway.

Enlarge / The pairing procedure for one of Eken’s doorbell cameras, which allows a malicious actor quite a bit of leeway.

Eken

Among the camera’s vulnerabilities cited by CR:

  • Sending public IP addresses and Wi-Fi SSIDs (names) over the Internet without encryption
  • Takeover of the cameras by putting them into pairing mode (which you can do from a front-facing button on some models) and connecting through the Aiwit app
  • Access to still images from the video feed and other information by knowing the camera’s serial number.

CR also noted that Eken cameras lacked an FCC registration code. More than 4,200 were sold in January 2024, according to CR, and often held an Amazon “Overall Pick” label (as one model did when an Ars writer looked on Wednesday).

“These video doorbells from little known manufacturers have serious security and privacy vulnerabilities, and now they’ve found their way onto major digital marketplaces such as Amazon and Walmart,” said Justin Brookman, director of tech policy at Consumer Reports, in a statement. “Both the manufacturers and platforms that sell the doorbells have a responsibility to ensure that these products are not putting consumers in harm’s way.”

CR noted that it contacted vendors where it found the doorbells for sale. Temu told CR that it would halt sales of the doorbells, but “similar-looking if not identical doorbells remained on the site,” CR noted.

A Walmart representative told Ars that all cameras mentioned by Consumer Reports, sold by third parties, have been removed from Walmart by now. The representative added that customers may be eligible for refunds and that Walmart prohibits the selling of devices that require an FCC ID and lack one.

Ars contacted Amazon for comment and will update this post with new information. An email sent to the sole address that could be found on Eken’s website was returned undeliverable. The company’s social media accounts were last updated at least three years prior.

Consumer Reports' researchers claim to have found JPEG file references passed in plaintext over the network, which could later be viewed without authentication in a browser.

Consumer Reports’ researchers claim to have found JPEG file references passed in plaintext over the network, which could later be viewed without authentication in a browser.

Consumer Reports

CR issued vulnerability disclosures to Eken and Tuck regarding its findings. The disclosures note the amount of data that is sent over the network without authentication, including JPEG files, the local SSID, and external IP address. It notes that after a malicious user has re-paired a doorbell with a QR code generated by the Aiwit app, they have complete control over the device until a user sees an email from Eken and reclaims the doorbell.

With a few exceptions, video doorbells and other IoT cameras tend to rely on cloud connections to stream and store footage, as well as notify their owners about events. This has led to some notable privacy and security concerns. Ring doorbells were found to be pushing Wi-Fi credentials in plaintext in late 2019. Eufy, a company that marketed its “No clouds” offerings, was found to be uploading facial thumbnails to cloud servers to send push alerts and later apologized for that and other vulnerabilities. Camera provider Wyze recently disclosed that, for the second time in five months, images and video feeds were accidentally available to the wrong customers following a lengthy outage.

Listing image by Amazon/Eken

$30 doorbell cameras have multiple serious security flaws, says Consumer Reports Read More »