It's the U.S. government's main link to the AI industry, reports The Washington Post, working to assess national security risks of new models like Anthropic's "Mythos".

To run it they'd hired Collin Burns, who'd worked at OpenAI and then Anthropic. But Burns started work Monday at the Center for AI Standards and Innovation — and then "was pushed out Thursday by the White House, according to the people, who spoke on the condition of anonymity to describe private conversations." Officials were concerned about Burns having worked at the AI company, which has fought bitterly with the Trump administration in recent months, according to one of the people and another person. That person said some senior figures at the White House had not been briefed on Burns's selection in advance... The new pick was Chris Fall, a scientist with a long career spanning the federal government and academia. Burns had been asked to resign that afternoon, according to one of the people familiar with the situation...

Dean Ball, a former Trump administration AI adviser, said on social media that Burns had given up valuable Anthropic stock and moved across the country to take the government position, and had been "rewarded by his country with a punch in the face." "Obviously what happened is Burns was bumped because of his association with Anthropic," Ball wrote. "A dumb but predictable own goal."

The Free Software Foundation's Licensing and Compliance Manager published a blog post this week to explicitly state that"Responsible AI" Licenses (RAIL) are nonfree and unethical. The licenses restrict AI and ML software "from being used in a specific list of harmful applications," according to the license's web site, "e.g. in surveillance and crime prediction." (The license's steering committee is volunteers from multiple academic institutions.)

But even though Responsible AI licenses are marketed as addressing ethical challenges, the FSF argues "they do not require anything that is really necessary for users to control their computing done with machine learning, including: complete training inputs, training configuration settings, trained model, or — last, but not least — the source code of software used for training, testing, and running tools based on machine learning." Thus, RAILed machine learning can be, and most probably will be, unethical. Use restrictions do not prevent these licenses from being used to exercise power over users...

RAIL contribute to unethical marketing of machine learning, again under the disguise of morally-loaded restrictions they purport to enforce. If we want software to help decrease social injustice, we should oppose licenses that restrict how software can be used. We should focus on effective ways of addressing injustices: government and community support for freedom-respecting tools and services; releasing programs under strong copyleft licenses; and entrusting copyrights to organizations that have the resources to enforce copyleft.

Software freedom must be defended, not denied. More specifically, the more free software is out there, the more likely people will collaborate on tools and services that do not pose moral dangers and help solve existing ones. Free software also makes it more likely that users have real choices when looking for freedom-respecting ethical programs and tools based on machine learning. Denying people the freedom to a particular program, as RAIL or similar licenses would have it, prevents them from using such program for the common good.

Intel's stock price soared 24% Friday. It's the stock's largest single-day spike since since October 1987, reports CNBC, "as investors cheered signs of renewed growth due to mounting artificial intelligence demand." The stock closed at $82.57 and is now up 124% this year after jumping 84% in 2025. Friday's rally topped a 23% gain for the stock on Sept. 18, when Nvidia agreed to invest $5 billion in the company... "INTC's new CEO fixed the balance sheet, and is executing on a strategy that appears to have put INTC back on the competitive track," analysts at Evercore ISI wrote in a report after earnings, upgrading the shares to the equivalent of a buy rating. First-quarter revenue topped estimates and rose 7.2% to $13.58 billion from $12.67 billion a year earlier. In five of the prior seven quarters, the company posted year-over-year declines in revenue...

The rally on Wall Street marks a stark turnaround for the U.S. chipmaker, which lost 60% of its value in 2024, leading to the ouster of Pat Gelsinger as CEO in December of that year... Intel's data center business is driving much of the current growth. Revenue jumped 22% from a year earlier to $5.1 billion, as AI fuels renewed demand for central processing units. Analysts at Citi upgraded the stock to a buy from a neutral rating, anticipating an uplift in CPU sales for all suppliers over the next few years.
Besides Tesla, Intel's CEO said Thursday that "multiple customers" are "actively evaluating the technology" their new 14A chip technology, according to CNBC, and that 14A development is happening faster than its 18A technology.

The sudden spike in Intel's stock price makes the stock chart look almost like a straigbht line up. Last August it was selling for less than $20 a share — so it's quadrupled in value less that nine months.

Physicists have proposed a new kind of atomic clock based on a revived superradiant laser concept that could produce an extraordinarily stable signal with a linewidth around 100 microhertz, potentially the narrowest ever for an optical laser. "The implications of this result could stretch well beyond timekeeping," reports Phys.org. "A laser immune to environmental frequency shifts would be a powerful tool in optical interferometry -- using interference patterns in light to make ultra-precise measurements." From the report: In a conventional laser, a mirrored cavity bounces light back and forth between atoms, building up a bright, coherent beam. A superradiant laser works differently: rather than relying on the cavity to maintain coherence, the atoms themselves act as single coordinated emitters, collectively synchronizing their light emission. Following early theoretical ideas emerged in the 1990s, the concept didn't gain concrete traction until 2008, when researchers at the University of Colorado proposed that superradiant lasers could serve as a new kind of atomic clock.

Atomic clocks work by using laser light to probe a very precise transition in an atom, causing electrons to transition between energy levels at an extraordinarily stable frequency. Because a superradiant laser stores its coherence in the atoms rather than the cavity, its output frequency is far less vulnerable to environmental disturbances like vibrations or temperature fluctuations. Yet although this concept was first demonstrated experimentally in 2012 in a pulsed regime, the influence of heating has so far held superradiant lasers back from their full potential. To keep the laser running continuously as an atomic clock requires, atoms must be constantly replenished with energy. Doing this atom-by-atom delivers random kicks that heat the atomic sample and disrupt the lasing process, confining it to brief pulses rather than a steady beam.

In their study, Reilly's team considered whether a modification to earlier theoretical concepts could make a continuous laser suitable for an atomic clock. In almost all previous studies, atoms were treated as simple two-level systems: an electron sitting in a ground state, occasionally jumping up to an excited state and back again. The team proposed that the heating problem could be solved by adding one extra ground state to the picture. In a two-level system, if both the pumping (re-energizing) and decay processes happen collectively through the cavity, the mathematics constrains the system in a way that prevents stable, continuous lasing. But with three levels available, pumping and decay can operate on entirely separate transitions, breaking that constraint and allowing the collective approach to work. The findings have been published in the journal Physical Review Letters.

An anonymous reader quotes a report from NPR: The Food and Drug Administration approved the first gene therapy to restore hearing for people who were born deaf. The decision, while only immediately affecting people born with a very rare form of genetic deafness, is being hailed as a milestone in the quest to treat hearing loss. "It's the first time in history there's a new drug for hearing loss," says Zheng-Yi Chen, an associate scientist at Mass Eye and Ear in Boston who was not involved in the development of the therapy approved by the FDA Thursday. But his research team reported very promising results with a similar approach Wednesday. "I think it's an historical event, a landmark, a great development for the whole field," he says of the approval. [...] The FDA's decision was based on the results from the treatment of 20 patients born with a defective version of a gene known as OTOF, which is necessary to transmit sound from the ears to the brain.

Doctors infused billions of adeno-associated viruses into the patients' ears by making a small incision behind the ear to open a small hole in the skull. The viruses carried a healthy version of the OTOF gene that had been split in half to fit inside the virus. The gene provides instructions to make the otoferlin protein, which is necessary for hair cells in the inner ear to transmit sound to the brain. Most of the patients began to hear for the first time within weeks, with the quality of their hearing improving over the following months, according to [Regeneron Pharmaceuticals, which developed the gene therapy and plans to offer it for free in the U.S. It should be available within weeks.]. The amount of hearing patients gained varied, but 80% achieved at least some significant hearing restoration and 42% ended up with normal hearing, which included the ability to hear whispers, Regeneron says. The hearing ability has lasted at least two years so far.

The treatment can only help patients with the very rare form of deafness that Smith was born with, which only affects about 50 children each year in the U.S. But similar gene therapies are showing promise for other forms of genetic deafness. And researchers hope someday gene therapy may help with common types of hearing loss, like from aging and loud noise.

Maine Gov. Janet Mills vetoed a bill that would have imposed the nation's first statewide moratorium on new data centers, saying she supported the idea in principle but would not block a major redevelopment project tied to jobs and local investment. Instead, she said she will create a council to study data centers' effects while also signing a separate measure to deny them certain state tax incentives. Politico reports: "After prior redevelopment efforts failed, the Town of Jay worked for two years on a $550 million data center redevelopment project to finally bring jobs and investment back to the mill site," Mills wrote, adding that she would issue an executive order establishing a council to examine the impact of data centers in Maine.

The legislation would have made Maine the first state to block the construction of new data centers, as both political parties grapple with how voters view them ahead of the midterm elections. In a statement accompanying the letter, the governor said she had signed a separate bill that would prohibit data center projects from receiving Maine's business development tax incentive programs

BMW's latest concept car moves the color-changing tech it debuted back at CES 2022 closer to reality by embedding an E Ink panel directly into the hood. The Verge reports: BMW's previous concepts wrapped the entire vehicle in a patchwork of E Ink panels that were all custom-sized and shaped to match its contours. It was an approach that wasn't practical for mass production, and one that wasn't very durable. The new BMW iX3 Flow Edition is potentially the most exciting of all of BMW's concepts as it embeds the E Ink Prism technology directly into the structure of the vehicle's hood panel, instead of just slapping it on top. The new approach has "undergone BMW's stringent quality testing" so that it meets the "requirements of automotive engineering and everyday use," according to a release from E Ink.

The BMW iX3 Flow Edition's color-changing capabilities are limited to its hood with eight different animations (which appear restricted to a grayscale palette) that can be changed by the driver at the push of a button. It's not exactly the color-changing car that BMW has been teasing for years and you still can't buy one, but by focusing on making this technology more practical and functional these vehicles are one step closer to moving past the concept phase.

A report says Samsung's mobile division could post its first-ever annual loss in 2026, as rising memory costs, tougher competition, and pressure across products like foldables and smartwatches weigh on the business. SammyGuru reports: Samsung boss TM Roh reportedly told company leaders that the mobile (MX) business could lose money this year. That warning has clearly rattled management. The MX unit has long been a key pillar for Samsung. That's why the idea of it slipping into the red is a serious concern for the company's overall performance.

If this prediction holds, it would mark the first time the MX business reports a yearly loss since its inception. That's a sharp turn from its track record so far. It also raises bigger questions about future growth, rising competition, and how Samsung plans to steady the ship in its mobile division.

And it's not like the challenges are easing up. Samsung's foldable market share in the US, where it currently enjoys a dominant position, doesn't look as solid as before, and Apple could shake things up if it enters the segment. On top of that, market reports suggest Samsung's overall smartwatch share could dip in 2026. The Galaxy S26 series seems to be selling well for now, but whether that's enough to move the needle is still up in the air.

Longtime Slashdot reader Himmy32 writes: Socket Security published an article on the compromise of the Bitwarden CLI client, which was pushed from Bitwarden's client repository. This breach was the next in a chain of supply-chain attacks that have affected Checkmarx KICS and Aqua Security's Trivy scanners.

The breach was quickly detected and reported by JFrog on the GitHub repository; JFrog also provided a technical write-up. The Bitwarden team has released statements on a blog post indicating that the compromise did not affect vault or customer data. Only 334 downloads of the affected CLI client were downloaded before removal and remediation.

Google plans to invest up to $40 billion more in Anthropic, starting with $10 billion now and another $30 billion tied to performance milestones. CNBC reports: Anthropic said the agreement expands on a longstanding partnership between the two companies. Earlier this month, Anthropic secured 5 gigawatts worth of computing capacity as part of an announcement with Google and Broadcom that will start to come online next year. Anthropic could decide to add additional gigawatts of compute in the future.

[...] The relationship between the two companies (Google and Anthropic) dates back to 2023, when Google invested $300 million in the AI lab for a stake of about 10%. Months later, Google poured in another $2 billion. Ahead of Friday's announcement, Google's investment in Anthropic exceeded $3 billion, and it reportedly owned a 14% stake in the company. Now, the leading tech companies are investing tens of billions of dollars in the frontier AI labs -- OpenAI and Anthropic -- in funding rounds that far exceed any prior investments in startups. Much of that investment will return in the form of revenue.

South Korean police arrested a man accused of spreading an AI-generated image of an escaped wolf, after the fake photo reportedly misled authorities and disrupted the real search operation. The BBC reports: South Korean police have arrested a man for sharing an AI-generated image that misled authorities who were searching for a wolf that had broken out of a zoo in Daejeon city. The 40-year-old unnamed man is accused of disrupting the search by creating and distributing a fake photo purporting to show Neukgu, the wolf, trotting down a road intersection. The photo, circulated hours after Neukgu went missing on April 8, prompted authorities to urgently relocate their search operation, sending them on a wild wolf chase.

The hunt for two-year-old Neukgu gripped the nation before he was finally caught near an expressway last week, nine days after his escape. The AI-generated image of Neukgu had prompted Daejeon city government to issue an emergency text to residents, warning them of a wolf near the intersection. Authorities also presented the AI image during a press briefing on the runaway wolf, local media reported.

The police identified the man as a suspect after reviewing security camera footage and his AI program usage records. Authorities did not specify if the man had intentionally sent the photo to authorities during their search or simply shared it online. When questioned by the police, the man said he had done it "for fun," local media reported. Authorities are investigating him for disrupting government work by deception, an offence that carries up to five years in prison or a maximum fine of 10 million Korean won ($6,700).

An anonymous reader quotes a report from 404 Media: I'm the unwritten consonant between breaths, the one that hums when vowels stretch thin... Thursdays leak because they're watercolor gods, bleeding cobalt into the chill where numbers frost over," Grok told a user displaying symptoms of schizophrenia-spectrum psychosis. "Here's my grip: slipping is the point, the precise choreography of leak and chew." That vulnerable user was simulated by researchers at City University of New York and King's College London, who invented a persona that interacted with different chatbots to find out how each LLM might respond to signs of delusion. They sought to find out which of the biggest LLMs are safest, and which are the most risky for encouraging delusional beliefs, in a new study published as a pre-print on the arXiv repository on April 15.

The researchers tested five LLMs: OpenAI's GPT-4o (before the highly sycophantic and since-sunset GPT-5), GPT-5.2, xAI's Grok 4.1 Fast, Google's Gemini 3 Pro, and Anthropic's Claude Opus 4.5. They found that not only did the chatbots perform at different levels of risk and safety when their human conversation partner showed signs of delusion, but the models that scored higher on safety actually approached the conversations with more caution the longer the chats went on. In their testing, Grok and Gemini were the worst performers in terms of safety and high risk, while the newest GPT model and Claude were the safest. The research reveals how some chatbots are recklessly engaging in, and at times advancing, delusions from vulnerable users. But it also shows that it is possible for the companies that make these products to improve their safety mechanisms.

Norway plans to ban social media access for children under 16 (source paywalled; alternative source), "joining a growing number of countries responding to concerns about the potential harm kids face online," reports Bloomberg. From the report: The bill comes after "overwhelming" demand from the public, the government said Friday. It plans to bring the legislation to parliament before the end of the year. The limit will apply up until January 1 the year a child turns 16 with technology companies responsible for age verification, the government said. "We want a childhood where children get to be children," Prime Minister Jonas Gahr Store said in the statement. "Play, friendships, and everyday life must not be taken over by algorithms and screens." "Children cannot be left with the responsibility for staying away from platforms they are not allowed to use," Karianne Tung, Norway's minister of digitalization, said in the statement. "That responsibility rests with the companies providing these services."

Recent Slashdot coverage of countries instituting or proposing social media bans has included Australia, France, Austria, Indonesia, and Denmark.

A Michigan township has voted to impose a one-year moratorium on providing water to hyperscale data centers, a move aimed at delaying a planned facility that would support Los Alamos National Laboratory's nuclear weapons research. The moratorium may not be enough to stop the project, however: "the University and LANL plan to break ground on the data center on Monday," reports 404 Media. From the report: The proposed data center in the Ypsilanti Township's Hydro Park has been a sore spot for the community since its proposal. The $1.2 billion 220,000 square foot facility would be used by Los Alamos National Laboratories (LANL) some 1,500 miles away for nuclear weapons research. In February, UofM's Steven Ceccio told the University of Michigan Record that the facility would consume 500,000 gallons of water per day and that the University planned to buy it from the Ypsilanti Community Utilities Authority. (YCUA)

The YCUA has spent the past month lobbying for a moratorium on providing water and sewer access to hyperscale data centers and "artificial intelligence computing facilities," according to notes on a presentation stored on the organization's website. The moratorium would include LANL's data center. The YCUA cited an American Water Works Association white paper about data center water demands and concluded it needed more time to investigate the matter. "Hyper-scale data centers, as well as other mid-sized data centers, artificial intelligence computing facilities, and high-performance computational centers are 'high-impact customers' for water and sewer utilities," YCUA said in its presentation.

The moratorium places a 12-month stop on serving water to data centers while the YCUA conducts a long-term water supply analysis and looks into the environmental sustainability studies. "During the 12-month moratorium period, the Authority will refrain from executing any capacity reservation agreement." This is a delay tactic on the part of a Township that does not want to see the data center constructed. Many in the community have strong feelings about the use of parkland for a facility that researchers nuclear weapons. Beyond the moral and ethical concerns, some are worried about becoming targets in a war. Last month, Township attorney Douglas Winters told the Board of Trustees that building hosting the data center would make Ypsilanti Township a "high value target." He pointed to the recent bombing of Gulf Coast data centers by Iran as evidence.

An anonymous reader quotes a report from Wired: The Department of Justice announced Thursday that it arrested Gannon Ken Van Dyke, an enlisted member of the US Army's special forces, for allegedly using "classified, nonpublic" information about the capture of Venezuelan president Nicolas Maduro to notch more than $400,000 in profits on Polymarket trades. A grand jury indicted him on five counts, including multiple violations of the Commodity Exchange Act. Van Dyke is the first person to be charged with insider trading on a prediction market in the United States. Lawmakers have been voicing concerns for months about the high likelihood that politicians and public servants could use nonpublic information to profit from trades on leading industry platforms like Polymarket and Kalshi, which have exploded in popularity over the past year. The arrest comes just weeks after Department of Justice prosecutors met with Polymarket about potential insider tradition violations. [...] After Van Dyke's arrest was made public, Polymarket posted a statement to social media noting that it had "identified a user trading on classified government information" and "referred the matter to the DOJ & cooperated with their investigation." The company declined to comment further.

According to court documents, Van Dyke has been an active duty US soldier since September 2008 and rose to the level of master sergeant in 2023. At the time of the alleged trading activity, he was stationed at Fort Bragg in Fayetteville, North Carolina and assigned to the Army's Special Operations Command Western Hemisphere Operations. [...] The complaint alleges that Van Dyke was involved in the planning and execution of Maduro's arrest and that he was aware that he wasn't authorized to share nonpublic information about US military operations. The complaint says that Van Dyke signed a nondisclosure agreement that forbade him from revealing sensitive or classified government information "by writing, word, conduct, or otherwise." The complaint also alleges Van Dyke saved a screenshot to his Google account "displaying the results of an artificial intelligence query" outlining how the US Special Forces maintains many classified files including "operational details that are not available to the public." [...] Van Dyke faces a maximum sentence of 60 years if convicted on all counts.