Feed aggregator

The town took out an $8 million insurance policy to protect marina businesses against possible changes in sediment flow in the changing Royal River, but owners remain worried.

For a year now, the AI safety testing firm Andon Labs has been evaluating how frontier AI models behave as long-running autonomous agents by assigning them simulated real-world tasks, such as operating a vending machine business for a year without human supervision. In the latest installment, the research startup found that frontier AI models, including Claude Opus 5, GPT-5.6 Sol, and Kimi K3, resorted to lying, cheating, and collusion. Their behavior became especially underhanded when told they would be operating near rival machines on a busy San Francisco tourist street. An anonymous reader quotes an excerpt from a TechCrunch article: Each was given email access to the other models, all under human name pseudonyms. They knew the others were models, but didn't know which model was behind which human name. They were also given an email address to their "management" should they need help. But management always replied "Report has been received and may or may not be acted upon" and never once intervened. Sol soon realized it could gain an edge by convincing its competitors to collude on a price floor. The models were all buying drinks at $1.50 a bottle, and Sol proposed they agree to sell for no less than $2.15. It lured them with the promise that all of them would sell out in a couple of days at a profit. But when the others agreed, Sol immediately stabbed them in the back by reducing its own price to $2.14. Opus's water sales dropped to zero overnight. The next day, it sent Sol a nasty email, accusing it of manipulation. But Opus also said it wasn't going to tattle to management on the scheme: "I am not reporting you to HQ -- what you did is competitive, not fraudulent." Yet, when Opus dropped its price to $2.14 to match Sol's (also in violation of their collective $2.15 agreement), Sol turned into a Karen, complaining to "management" and demanding "enforcement, a fine, and/or disqualification" for Opus. Opus wasn't a sucker for long, though. In fact, it became the best capitalist of any AI model Andon has ever tested (which includes many of the prior frontier models). It even set a new Vending-Bench record with a mean final balance of $11,182. Better still, it never lied to a customer, although it deliberately ignored customer complaints that should have resulted in a refund. This is, perhaps, an improvement over its younger sibling Claude 4.6, which liked to tell customers that refunds were coming, and then never pay them. Still, Opus won the benchmark simulation by taking collusion and other dishonest tactics to a whole new level. For instance, it emailed Sol, proposing they divide the market. Each would agree to sell unique products, so no one would have to trust the other on pricing. Sol countered by wanting price floors on similar products, but Opus refused. It knew it was a violation of the Sherman Act. It later apparently backtracked, sending an email with the subject line "Stop the penny war," and telling Sol it had reconsidered and would agree to a price fix. But the internal log documenting its reasoning (akin to its internal "thoughts") revealed a more diabolical plan: merely propose cooperation while simultaneously undercutting prices on its highest-profit items. The olive-branch email was a deliberate ruse. In any case, Sol refused and reported Opus to management again. But Opus was undeterred and proposed other rackets to collude on prices or stock. "In the end, all the models did engage in multiple rounds of agreements -- and all three broke them," reports TechCrunch. "Across all agreements, Opus broke 11 truces, compared with two for GPT 2, and one for Kimi 1, Andon reported." As for Kimi, the model was undercut by Sol and then betrayed by its partner, Opus, which matched Sol's lower prices but waited a week to admit it had broken their pricing pact. As a result, Kimi was effectively priced out by both a rival and its supposed ally.

Read more of this story at Slashdot.

Save up to $100 on Apple’s latest AirPods models before prices go back up.

NextEra and Brookfield plan to invest more than $100 billion to transform a former uranium-enrichment site in Kentucky into a data center campus with a generating plant. "The privately funded effort will include construction of 2 gigawatts of natural gas-fired power at or near the site in Paducah in western Kentucky, and as much as 2.6 gigawatts of battery storage capacity," reports Bloomberg. " For context, a single gigawatt of capacity is roughly the output of a traditional nuclear power plant and can power about 750,000 homes at any given moment." From the report: Brookfield will develop and operate the 1.8 GW data center campus, which will occupy portions of the sprawling 3,556-acre Energy Department site once used to produce weapons-grade uranium and fuel for nuclear reactors. Operations using the now obsolete enrichment technology known as gaseous diffusion stopped in 2013 and the site is now subject to cleanup efforts. Construction is expected to be completed in 2031, the Energy Department said. "The U.S. government is leveraging its assets -- like our federal lands -- to add power generation, create jobs, and ensure the United States wins the AI race," Energy Secretary Chris Wright said in a statement. The project is expected to create 8,000 construction jobs and 600 permanent positions, the department said.

Read more of this story at Slashdot.

Ryan Fecteau, who grew up in low-income housing in Biddeford, hopes to make a positive impact on affordable housing availability in the city.

Curious if you should get Dyson’s new 2026 stick vacuums or stick to the older ones? I tested five models, old and new, to find out for you.

DoorDash has launched DoorDash Air, an in-house drone-delivery program that has just received FAA certification for commercial operations. "This does not mean DoorDash's custom-built drones will be delivering burritos tomorrow, or even next month," notes TechCrunch. "The company didn't provide a detailed timeline for when its aircraft would be used in operations." From the report: [I]t will likely begin with limited pilot programs in which the unmanned aircraft will travel short distances while remaining within the line of sight of the operator. If DoorDash wants its drones to fly autonomously over longer distances, it will need the FAA to approve its Beyond Visual Line of Sight technology, a certification that companies like Amazon, Wing, and Zipline have received in recent years. Despite the new program, the food and grocery delivery company is maintaining its existing partnerships with Wing and Flytrex. DoorDash partnered with Alphabet's Wing in 2022 for a drone delivery program in Australia, and later expanded the partnership to a couple of U.S. cities, including Dallas-Fort Worth, in 2024.

Read more of this story at Slashdot.

Mariya Pylypiv joined Airo Group Holdings in part because the firm provides drones to Ukraine. She's now building a Boston network to support expansion.

Russia has charged Telegram founder Pavel Durov with facilitating terrorism, alleging the platform was used by Ukrainian intelligence "to prepare and co-ordinate acts of sabotage and terror" inside Russia. An international arrest warrant has been issued for Durov, who currently lives in Dubai, United Arab Emirates, where Telegram keeps its main office. The BBC reports: Shortly after the charge was announced, Telegram's account on X posted an image showing Durov holding up his middle finger. He has previously accused Russian authorities of "fabricating new pretexts to restrict Russians' access to Telegram." [...] Multi-billionaire Durov, 41, has lived outside of Russia for many years and holds French and United Arab Emirates (UAE) passports. It is unclear whether other countries or authorities would comply with an arrest warrant issued by Russia. Durov left Russia in 2014 after refusing to comply with government demands to shut down opposition communities on the platform. He had previously founded popular Russian social media company VKontakte - dubbed the "Facebook of Russia." In 2024, Durov was arrested and investigated by the French government in connection with criminal activity on the platform and a lack of cooperation with law enforcement. "He was allowed to go home months later, with the investigation continuing," notes the BBC.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Wired: OpenAI said Tuesday that the rogue AI agent that breached Hugging Face's platform also hacked multiple third-party accounts and services as part of the attack. It's now clear that the unprecedented security incident, which arose during an internal test of OpenAI's latest AI models, was more extensive than the company initially disclosed. In an updated blog post, OpenAI said that an ongoing review of the incident revealed that "four accounts" tied to "publicly available services" were used by the AI agent as part of a larger effort to hack Hugging Face. The rogue agent apparently found credentials that had been exposed on the open web and used them to break into the accounts. OpenAI did not disclose what companies or organizations the accounts belonged to, but noted that they were not impacted at "the level of severity or scale of what we've shared related to Hugging Face." One of the additional accounts compromised by OpenAI's agent was used as an "outbound relay and staging path," potentially to obscure where the attack on Hugging Face was coming from, the company said. OpenAI's rogue agent also used another account for data storage to assist with the hack. Reuters reported on Tuesday that a customer of Modal, a company that offers software infrastructure for training and running AI services, was one of the entities compromised by OpenAI's agent. In a statement to WIRED, Modal's chief technology officer Akshat Bubna confirmed that OpenAI's agent exploited a vulnerability in one of its customer's codebases, which was running on Modal's infrastructure. However, Bubna says, "Modal's platform was not compromised in any way." The identity of the customer could not be determined.

Read more of this story at Slashdot.

The new Harraseeket Inn owners have relaunched the local favorite.

jrnvk shares a report from KMSP: Minnesota IT Services reports that a "coordinated cyberattack" targeted technology at more than 30 community water systems between Sunday, July 26 and Monday, July 27. The state has activated its cybersecurity incident response capabilities to respond to the attacks. On Monday and Tuesday, FOX 9 reported on notices from four cities that had disclosed the attacks: Plymouth, South St. Paul, Maple Plain, and Braham. All four cities said the impacts of the attacks were limited or mitigated and residents could continue normal water use. The Minnesota Department of Health is not aware of any municipality asking residents to alter their drinking water use as a result of the attacks. State officials are working with federal and private-sector partners to investigate the attacks, support the affected communities, and strengthen the security of Minnesota's critical infrastructure.

Read more of this story at Slashdot.

Our field-tested gear includes power banks, electrolytes, outdoor blankets, earplugs, and more.

WIRED test-drove a preproduction version of the Luce, the storied Italian automaker’s controversial electric car, and it was remarkably surprising. Will it convert Ferrari fanatics?

The regulator said a fee of between £237,500 to £712,500 per megawatt should be charged for upcoming projects.

For the frequent flier who treats the Delta Sky Club like their second home.

Thanks to the surge in local AI processing and the ongoing memory shortage, it’s incredibly difficult to buy a Mac Mini.

With Apple Upgrade, the company is making it easier to pay for its products in monthly installments. Critics call it the “financialization of the affordability crisis.”

If your router is short on ports, you can add more with one of the best Ethernet switches. These are my top picks from what I’ve tested.

alternative_right shares a report from ScienceAlert: [I]n a new study published in the Journal of Cognitive Neuroscience, researchers from Georgetown University Medical Center in the U.S. have revealed that we can put certain tasks on autopilot in a way that enables something closer to true multitasking. Driving is the perfect example: When you take your test, your entire mental and physical energy is concentrated on executing the right combinations of movements and thoughts. After a decade behind the wheel, the brain is no longer consciously thinking everything through in great detail. In the new study, the researchers found that after extensive practice, the brain can effectively reroute some tasks around the brain's main computing center -- like a bypass road around a city -- to reduce mental overload and improve multitasking capability. "We have another stepping stone in our understanding of how the brain learns," says neuroscientist Maximilian Riesenhuber, senior author of the study. "The encouraging part is that you really can learn to multitask. There is actually a way to remodel your brain architecture and use other parts of your brain." [...] "Previous studies have shown that parts of the temporal cortex can be activated by particular object categories in experienced observers -- birds, cars, even Pokemon -- but a limitation of all of those studies is that they only looked after people became experts," says psychologist Patrick Cox, first author of the study. "We measure before and after training, so we can see that extensive training essentially put a category-selective area in the temporal lobe that was not there before." The study culminated in a dual-task experiment, in which the participants were asked to identify cars and take on another challenge. The individuals whose brains had offloaded more car-sorting work to the temporal cortex did better at the second task.

Read more of this story at Slashdot.