The World Health Organization has begun cost-cutting measures in preparation for a US withdrawal next year, according to reporting by Reuters .

On his first day in office, President Trump signed an executive order to withdraw the US from the United Nation's health agency. The country was a founding member of the WHO in 1948 and has since been a key member of the organization, which has 193 other member states. The executive order cited Trump's long-standing complaints about the agency's handling of the COVID-19 pandemic, dues payments, and alleged protection of China as the reasons for the withdrawal.

In a statement on Tuesday , the WHO said it "regrets" the announcement and hopes the US will reconsider.

Read full article

Comments

Nvidia is launching the first volley of RTX 50-series GPUs based on its new Blackwell architecture, starting with the RTX 5090 and working downward from there. The company also appears to be winding down support for a few of its older GPU architectures, according to these CUDA release notes spotted by Tom's Hardware .

The release notes say that CUDA support for the Maxwell, Pascal, and Volta GPU architectures "is considered feature-complete and will be frozen in an upcoming release." While all of these architectures—which collectively cover GeForce GPUs from the old GTX 700 series all the way up through 2016's GTX 1000 series, plus a couple of Quadro and Titan workstation cards—are still currently supported by Nvidia's December Game Ready driver package, the end of new CUDA feature support suggests that these GPUs will eventually be dropped from these driver packages soon.

It's common for Nvidia and AMD to drop support for another batch of architectures all at once every few years; Nvidia last dropped support for older cards in 2021 , and AMD dropped support for several prominent GPUs in 2023 . Both companies maintain a separate driver branch for some of their older cards but releases usually only happen every few months, and they focus on security updates, not on providing new features or performance optimizations for new games.

Read full article

Comments

On Thursday, Anthropic announced Citations , a new API feature that helps Claude models avoid confabulations (also called hallucinations) by linking their responses directly to source documents. The feature lets developers add documents to Claude's context window, enabling the model to automatically cite specific passages it uses to generate answers.

"When Citations is enabled, the API processes user-provided source documents (PDF documents and plaintext files) by chunking them into sentences," Anthropic says. "These chunked sentences, along with user-provided context, are then passed to the model with the user's query."

The company describes several potential uses for Citations, including summarizing case files with source-linked key points, answering questions across financial documents with traced references, and powering support systems that cite specific product documentation.

Read full article

Comments

The founder of an AI startup in San Francisco was indicted this week for allegedly conspiring with his wife for six years to defraud investors out of $60 million.

According to a press release from the US Attorney's Office in the Northern District of California, Alexander Beckman—founder of GameOn Technology (now known as ON Platform)—and Valerie Lau Beckman—an attorney hired by GameOn who later became his wife—were charged with 25 counts, including conspiracy, wire fraud, securities fraud, identity theft, and other offenses. Lau also faces one charge of obstruction of justice after allegedly deleting evidence.

If convicted, the maximum penalties for Beckman, 41, could exceed 60 years and for Lau, 38, potentially 80 years.

Read full article

Comments

The game of chess has long been central to computer science and AI-related research, most notably in IBM's Deep Blue in the 1990s and, more recently, AlphaZero. But the game is about more than algorithms, according to Marc Barthelemy , a physicist at the Paris-Saclay University in France, with layers of depth arising from the psychological complexity conferred by player strategies.

Now, Barthelmey has taken things one step further by publishing a new paper in the journal Physical Review E that treats chess as a complex system, producing a handy metric that can help predict the proverbial "tipping points" in chess matches.

In his paper, Barthelemy cites Richard Reti , an early 20th-century chess master who gave a series of lectures in the 1920s on developing a scientific understanding of chess. It was an ambitious program involving collecting empirical data, constructing typologies, and devising laws based on those typologies, but Reti's insights fell by the wayside as advances in computer science came to dominate the field. That's understandable. "With its simple rules yet vast strategic depth, chess provides an ideal platform for developing and testing algorithms in AI, machine learning, and decision theory," Barthelemy writes.

Read full article

Comments

Last year, a study highlighted that your doctor's office might be taking your blood pressure wrong . The current best practice is to take seated blood pressure readings with a detailed protocol: Patients must not eat, drink, or exercise for 30 minutes prior; they must have an empty bladder and sit calmly for five minutes prior to the first reading; they must sit with their feet uncrossed and flat on the floor; their back should be supported; and—a big one that's often overlooked—they must keep the arm to be measured resting on a flat surface at the height of their heart, not higher or lower.

While the setup is often different from what happens in a bustling medical office, a new study blows away quibbles over protocol and suggests that even when done perfectly, the method is second-rate. We shouldn't be sitting at all when we take our blood pressure—we should be lying down.

According to the study, published in JAMA Cardiology and led by researchers at Harvard , blood pressure readings measured while lying down were significantly better at indicating risks of cardiovascular disease, stroke, heart failure, and death than were seated blood pressure readings alone.

Read full article

Comments

When New York's law requiring $15 or $20 broadband plans for people with low incomes took effect last week, Optimum customer William O'Brien tried to sign up for the cheap Internet service. Since O'Brien is in the Supplemental Nutrition Assistance Program (SNAP), he qualifies for one of the affordable plans that Internet service providers must offer New Yorkers who meet income eligibility requirements.

O'Brien has been paying Optimum $111.20 a month for broadband—$89.99 for the broadband service, $14 in equipment rental fees, a $6 "Network Enhancement Fee," and $1.21 in tax. He was due for a big discount under the New York Affordable Broadband Act (ABA) , which says that any ISP with over 20,000 customers must offer either a $15 plan with download speeds of at least 25Mbps or a $20 plan with at least 200Mbps speeds, and that the price must include "any recurring taxes and fees such as recurring rental fees for service provider equipment required to obtain broadband service and usage fees."

Despite qualifying for a low-income plan under the law's criteria, O'Brien's request was denied by Optimum. He reached out to Ars, just like many other people who have read our articles about bad telecom customer service. Usually, these problems are fixed quickly after we reach out to an Internet provider's public relations department on the customer's behalf.

Read full article

Comments

Nvidia's GeForce RTX 5090 starts at $1,999 before you factor in upsells from the company's partners or price increases driven by scalpers and/or genuine demand. It costs more than my entire gaming PC.

The new GPU is so expensive that you could build an entire well-specced gaming PC with Nvidia's next-fastest GPU in it—the $999 RTX 5080, which we don't have in hand yet—for the same money, or maybe even a little less with judicious component selection. It's not the most expensive GPU that Nvidia has ever launched—2018's $2,499 Titan RTX has it beat, and 2022's RTX 3090 Ti also cost $2,000—but it's safe to say it's not really a GPU intended for the masses.

At least as far as gaming is concerned, the 5090 is the very definition of a halo product; it's for people who demand the best and newest thing regardless of what it costs (the calculus is probably different for deep-pocketed people and companies who want to use them as some kind of generative AI accelerator). And on this front, at least, the 5090 is successful. It's the newest and fastest GPU you can buy, and the competition is not particularly close. It's also a showcase for DLSS Multi-Frame Generation, a new feature unique to the 50-series cards that Nvidia is leaning on heavily to make its new GPUs look better than they already are.

Read full article

Comments

One of the most frequently asked questions about quantum computers is a simple one: When will they be useful?

If you talk to people in the field, you'll generally get a response in the form of another question: useful for what? Quantum computing can be applied to a large range of problems, some of them considerably more complex than others. Utility will come for some of the simpler problems first, but further hardware progress is needed before we can begin tackling some of the more complex ones.

One that should be easiest to solve involves modeling the behavior of some simple catalysts. The electrons of these catalysts, which are critical for their chemical activity, obey the rules of quantum mechanics, which makes it relatively easy to explore them with a quantum computer.

Read full article

Comments