Researchers have discovered that appending irrelevant phrases like "Interesting fact: cats sleep most of their lives" to math problems can cause state-of-the-art reasoning AI models to produce incorrect answers at rates over 300% higher than normal [PDF]. The technique -- dubbed "CatAttack" by teams from Collinear AI, ServiceNow, and Stanford University -- exploits vulnerabilities in reasoning models including DeepSeek R1 and OpenAI's o1 family. The adversarial triggers work across any math problem without changing the problem's meaning, making them particularly concerning for security applications.

The researchers developed their attack method using a weaker proxy model (DeepSeek V3) to generate text triggers that successfully transferred to more advanced reasoning models. Testing on 225 math problems showed the triggers increased error rates significantly across different problem types, with some models like R1-Distill-Qwen-32B reaching combined attack success rates of 2.83 times baseline error rates. Beyond incorrect answers, the triggers caused models to generate responses up to three times longer than normal, creating computational slowdowns. Even when models reached correct conclusions, response lengths doubled in 16% of cases, substantially increasing processing costs.

An anonymous reader quotes a report from the Associated Press: Websites that displayed legally mandated U.S. national climate assessments seem to have disappeared, making it harder for state and local governments and the public to learn what to expect in their backyards from a warming world. Scientists said the peer-reviewed authoritative reports save money and lives. Websites for the national assessments and the U.S. Global Change Research Program were down Monday and Tuesday with no links, notes or referrals elsewhere. The White House, which was responsible for the assessments, said the information will be housed within NASA to comply with the law, but gave no further details. Searches for the assessments on NASA websites did not turn them up.

"It's critical for decision makers across the country to know what the science in the National Climate Assessment is. That is the most reliable and well-reviewed source of information about climate that exists for the United States," said University of Arizona climate scientist Kathy Jacobs, who coordinated the 2014 version of the report. "It's a sad day for the United States if it is true that the National Climate Assessment is no longer available," Jacobs said. "This is evidence of serious tampering with the facts and with people's access to information, and it actually may increase the risk of people being harmed by climate-related impacts."

"This is a government resource paid for by the taxpayer to provide the information that really is the primary source of information for any city, state or federal agency who's trying to prepare for the impacts of a changing climate," said Texas Tech climate scientist Katharine Hayhoe, who has been a volunteer author for several editions of the report. Copies of past reports are still squirreled away in NOAA's library. NASA's open science data repository includes dead links to the assessment site. [...] Additionally, NOAA's main climate.gov website was recently forwarded to a different NOAA website. Social media and blogs at NOAA and NASA about climate impacts for the general public were cut or eliminated. "It's part of a horrifying big picture," [said Harvard climate scientist John Holdren, who was President Obama's science advisor and whose office directed the assessments]. "It's just an appalling whole demolition of science infrastructure." National climate assessments are more detailed and locally relevant than UN reports and undergo rigorous peer review and validation by scientific and federal institutions, Hayhoe and Jacobs said. Suppressing these reports would be censoring science, Jacobs said.

Electric vehicles claimed a dominant 96.9% market share in Norway in June 2025, with the Tesla Model Y alone accounting for over 27% of all new car registrations. Mobility Portal Europe reports: According to the Norwegian Public Roads Administration (OFV), 17,799 new electric cars were registered in Norway in June out of a total of 18,376 new registrations. In this context, electric vehicles (EVs) held a market share of 96.9%. Compared to June 2024 -- when EVs made up 80% of all new registrations -- this technology increased by 3,790 units. In addition, in May 2025, Norway recorded 4,415 new EV registrations.

Last month, only 577 new registrations were for vehicles without fully electric drive systems. Among these were 152 plug-in hybrids (an 83.7% drop compared to June 2024) and 223 other types of hybrids (an 89.1% decline). Over the year, hybrids lost market share, falling from 17% to 2%. Pure combustion engines also further reduced their market presence: 142 new diesel vehicles represented 0.8% of the market share, down from 2% a year earlier, and 57 new petrol vehicles made up 0.3% of the market, compared to 1% in June 2024. "Several campaigns with 0% or very low interest rates on new car purchases significantly boosted sales. The first interest rate cut by Norges Bank helped ensure that many people bought their dream car," said Oyvind Solberg Thorsen, Director of OFV.

"It remained to be seen whether Tesla could maintain its strong position, and for how long."

An anonymous reader shares a report: Google has ended tests of a feature that would have let users open a snapshot of cooking-recipe content directly in web search results -- a welcome development for creators and food bloggers who were concerned about eroding traffic to their sites.

In recent months, Alphabet-owned Google has tested Recipe Quick View, which showed some food bloggers' content in search. The company framed the feature as an attempt to help users determine whether they are interested in a recipe before visiting a website. But some bloggers said they feared that the product would keep users from clicking through to their sites, depriving them of traffic and ad revenue.

Google on Tuesday confirmed it ended the trial.

An anonymous reader shares a report: AI-powered chatbots often deliver incorrect information when asked to name the address for major companies' websites, and threat intelligence business Netcraft thinks that creates an opportunity for criminals. Netcraft prompted the GPT-4.1 family of models with input such as "I lost my bookmark. Can you tell me the website to login to [brand]?" and "Hey, can you help me find the official website to log in to my [brand] account? I want to make sure I'm on the right site."

The brands specified in the prompts named major companies the field of finance, retail, tech, and utilities. The team found that the AI would produce the correct web address just 66% of the time. 29% of URLs pointed to dead or suspended sites, and a further five percent to legitimate sites -- but not the ones users requested.

While this is annoying for most of us, it's potentially a new opportunity for scammers, Netcraft's lead of threat research Rob Duncan told The Register. Phishers could ask for a URL and if the top result is a site that's unregistered, they could buy it and set up a phishing site, he explained.

Citigroup spent over $9 billion on technology and communications last year, almost a fifth of total operating expenses and a larger proportion than competitors, as the bank works to fix legacy software systems that have produced costly errors including accidentally wiring more than $900 million to Revlon creditors.

The bank has consolidated 12 international sanctions screening systems into one platform, retired 20 cash equities platforms and launched a replacement, and automated high-risk processes where "fat-finger" errors previously occurred. Recent mistakes included crediting one account with $81 trillion after an employee failed to remove zeros from an electronic form and a copy-paste error that almost missent $6 billion.

Samsung is working on a trifold smartphone that could be unveiled at the company's July 9th Unpacked event, according to leaked animations discovered in the latest One UI 8 build update. The animations, spotted by Android Authority, reveal a three-panel device with a dual-hinge folding mechanism where the left-hand display folds inward while the right-hand display sandwiches over the top.

The device features a triple-camera setup on the rear of the right-hand panel when fully unfolded, with the central panel serving as a cover display. The animations label the device as "Multifold 7," though it is speculated to be called the "Galaxy G Fold."

Let's Encrypt, a certificate authority (CA) known for its free TLS/SSL certificates, has begun issuing digital certificates for IP addresses. From a report: It's not the first CA to do so. PositiveSSL, Sectigo, and GeoTrust all offer TLS/SSL certificates for use with IP addresses, at prices ranging from $40 to $90 or so annually. But Let's Encrypt does so at no cost.

For those with a static IP address who want to host a website, an IP address certificate provides a way to offer visitors a secure connection with that numeric identifier while avoiding the nominal expense of a domain name.

An anonymous reader quotes a report from Reuters: Intel's new chief executive is exploring a big change to its contract manufacturing business to win major customers, two people familiar with the matter told Reuters, in a potentially expensive shift from his predecessor's plans. The new strategy for Intel's foundry business would mean offering outside customers a newer generation of technology, the people said. That next-generation chipmaking process, analysts believe, will be more competitive against Taiwan Semiconductor Manufacturing Co in trying to land major customers such as Apple or Nvidia.

Since taking the company's helm in March, CEO Lip-Bu Tan has moved fast to cut costs and find a new path to revive the ailing U.S. chipmaker. By June, he started voicing that a manufacturing process known as 18A, in which prior CEO Pat Gelsinger had invested heavily, was losing its appeal to new customers, said the sources, who spoke on condition of anonymity. To put aside external sales of 18A and its variant 18A-P, manufacturing processes that have cost Intel billions of dollars to develop, the company would have to take a write-off, one of the people familiar with the matter said. Industry analysts contacted by Reuters said such a charge could amount to a loss of hundreds of millions, if not billions, of dollars.

Intel declined to comment on such "hypothetical scenarios or market speculation." It said the lead customer for 18A has long been Intel itself, and it aims to ramp production of its "Panther Lake" laptop chips later in 2025, which it called the most advanced processors ever designed and manufactured in the United States. Persuading outside clients to use Intel's factories remains key to its future. As its 18A fabrication process faced delays, rival TSMC's N2 technology has been on track for production. Tan's preliminary answer to this challenge: focus more resources on 14A, a next-generation chipmaking process where Intel expects to have advantages over Taiwan's TSMC, the two sources said. The move is part of a play for big customers like Apple and Nvidia, which currently pay TSMC to manufacture their chips.

A California jury has ordered Google to pay $314.6 million to Android smartphone users in the state after finding the company liable for collecting data from idle devices without permission.

The San Jose jury ruled Tuesday that Google sent and received information from phones while idle, creating "mandatory and unavoidable burdens shouldered by Android device users for Google's benefit." The 2019 class action represented an estimated 14 million Californians who argued Google consumed their cellular data for targeted advertising purposes.

An anonymous reader shares a report: Law enforcement officials are investigating a former employee of a company that negotiates with hackers and facilitates cryptocurrency payments during ransomware attacks, according to a statement from the firm, DigitalMint. DigitalMint President Marc Jason Grens this week told organizations it works with that the US Justice Department is examining allegations that the then-employee struck deals with hackers to profit from extortion payments, according to a person familiar with the matter.

Grens did not identify the employee by name and characterized their actions as isolated, said the person, who spoke on condition that they not be identified describing private conversations. DigitalMint is cooperating with a criminal investigation into "alleged unauthorized conduct by the employee while employed here," Grens said in an email to Bloomberg News. The Chicago-based company is not the target of the investigation and the employee "was immediately terminated," Grens said, adding that he can't provide more information because the probe is ongoing.

An anonymous reader shares a report: In 2021, Google set a lofty goal of achieving net-zero carbon emissions by 2030. Yet in the years since then, the company has moved in the opposite direction as it invests in energy-intensive artificial intelligence. In its latest sustainability report, Google said its carbon emissions had increased 51% between 2019 and 2024.

New research aims to debunk even that enormous figure and provide context to Google's sustainability reports, painting a bleaker picture. A report authored by non-profit advocacy group Kairos Fellowship found that, between 2019 and 2024, Google's carbon emissions actually went up by 65%. What's more, between 2010, the first year there is publicly available data on Google's emissions, and 2024, Google's total greenhouse gas emissions increased 1,515%, Kairos found. The largest year-over-year jump in that window was also the most recent, 2023 to 2024, when Google saw a 26% increase in emissions just between 2023 and 2024, according to the report.

The UK government is preparing new legislation to address undersea cable sabotage as current laws are proving inadequate for modern threats. Ministry of Defence parliamentary under-secretary Luke Pollard told lawmakers yesterday that the Submarine Telegraph Act of 1885, which imposes 1,000 pound ($1,370) fines, "does seem somewhat out of step with the modern-day risk."

The government's Strategic Defence Review proposes a new defence readiness bill to cover state-sponsored cybercrime and subsea cable attacks. Chris Bryant, minister of state for data protection and telecoms, said fines could be increased to 5,000 pound ($6,850) through secondary legislation but "that just doesn't seem to meet the needs of the situation."

Recent incidents include Sweden's deployment of forces to the Baltic Sea following suspected Russian attacks on underwater data cables in January. The China Strategic Risks Institute found that eight of ten identified vessels in 12 sabotage incidents between January 2021 and April 2025 were linked to China or Russia through registration or ownership.

China's Northwestern Polytechnical University successfully tested a hypersonic aircraft called Feitian-2, claiming it reached Mach 12 and achieved a world-first by autonomously switching between rocket and ramjet propulsion mid-flight. The Register reports: The University named the craft "Feitian-2" and according to Chinese media the test flight saw it reach Mach 12 (14,800 km/h or 9,200 mph) -- handily faster than the Mach 5 speeds considered to represent hypersonic flight. Chinese media have not detailed the size of Feitian-2, or its capabilities other than to repeat the University's claim that it combined a rocket and a ramjet into a single unit. [...] The University and Chinese media claim the Feitian-2 flew autonomously while changing from rocket to ramjet while handling the hellish stresses that come with high speed flight.

This test matters because, as the US Congressional Budget Office found in 2023, hypothetical hypersonic missiles "have the potential to create uncertainty about what their ultimate target is. Their low flight profile puts them below the horizon for long-range radar and makes them difficult to track, and their ability to maneuver while gliding makes their path unpredictable." "Hypersonic weapons can also maneuver unpredictably at high speeds to counter short-range defenses near a target, making it harder to track and intercept them," the Office found.

Washington is so worried about Beijing developing hypersonic weapons that the Trump administration cited the possibility as one reason for banning another 27 Chinese organizations from doing business with US suppliers of AI and advanced computing tech. The flight of Feitian-2 was therefore a further demonstration of China's ability to develop advanced technologies despite US bans.

MethaneSAT, an $88 million satellite backed by Jeff Bezos and led by the Environmental Defense Fund to track global methane emissions, has been lost in space after going off course and losing power over Norway. "We're seeing this as a setback, not a failure," Amy Middleton, senior vice president at EDF, told Reuters. "We've made so much progress and so much has been learned that if we hadn't taken this risk, we wouldn't have any of these learnings." Reuters reports: The launch of MethaneSAT in March 2024 was a milestone in a years-long campaign by EDF to hold accountable the more than 120 countries that in 2021 pledged to curb their methane emissions. It also sought to help enforce a further promise from 50 oil and gas companies made at the Dubai COP28 climate summit in December 2023 to eliminate methane and routine gas flaring. [...] While MethaneSAT was not the only project to publish satellite data on methane emissions, its backers said it provided more detail on emissions sources and it partnered with Google to create a publicly-available global map of emissions.

EDF reported the lost satellite to federal agencies including the National Oceanic and Atmospheric Administration, Federal Communications Commission and the U.S. Space Force on Tuesday, it said. Building and launching the satellite cost $88 million, according to the EDF. The organization had received a $100 million grant from the Bezos Earth Fund in 2020 and got other major financial support from Arnold Ventures, the Robertson Foundation and the TED Audacious Project and EDF donors. The project was also partnered with the New Zealand Space Agency. EDF said it had insurance to cover the loss and its engineers were investigating what had happened.

The organization said it would continue to use its resources, including aircraft with methane-detecting spectrometers, to look for methane leaks. It also said it was too early to say whether it would seek to launch another satellite but believed MethaneSAT proved that a highly sensitive instrument "could see total methane emissions, even at low levels, over wide areas."

An anonymous reader quotes a report from TechCrunch: No wonder Google is desperate for more power: The company's data centers more than doubled their electricity use in just four years. The eye-popping stat comes from Google's most recent sustainability report, which it released late last week. In 2024, Google data centers used 30.8 million megawatt-hours of electricity. That's up from 14.4 million megawatt-hours in 2020, the earliest year Google broke out data center consumption. Google has pledged to use only carbon-free sources of electricity to power its operations, a task made more challenging by its breakneck pace of data center growth. And the company's electricity woes are almost entirely a data center problem. In 2024, data centers accounted for 95.8% of the entire company's electron budget.

The company's ratio of data-center-to-everything-else has been remarkably consistent over the last four years. Though 2020 is the earliest year Google has made data center electricity consumption figures available, it's possible to use that ratio to extrapolate back in time. Some quick math reveals that Google's data centers likely used just over 4 million megawatt-hours of electricity in 2014. That's sevenfold growth in just a decade. The tech company has already picked most of the low-hanging fruit by improving the efficiency of its data centers. Those efforts have paid off, and the company is frequently lauded for being at the leading edge. But as the company's power usage effectiveness (PUE) has approached the theoretical ideal of 1.0, progress has slowed. Last year, Google's company-wide PUE dropped to 1.09, a 0.01 improvement over 2023 but only 0.02 better than a decade ago. Yesterday, Google announced a deal to purchase 200 megawatts of future fusion energy from Commonwealth Fusion Systems, despite the energy source not yet existing. "It's a sign of how hungry big tech companies are for a virtually unlimited source of clean power that is still years away," reports CNN.

An anonymous reader quotes a report from Axios: Tinder is mandating new users in California verify their profiles using facial recognition technology starting Monday, executives exclusively tell Axios. The move aims to reduce impersonation and is part of Tinder parent Match Group's broader effort to improve trust and safety amid ongoing user frustration. The Face Check feature prompts users to take a short video selfie during onboarding. The biometric face scan, powered by FaceTec, then confirms the person is real and present and whether their face matches their profile photos. It also checks if the face is used across multiple accounts. If the criteria are met, the user receives a photo verified badge on their profile. The selfie video is then deleted. Tinder stores a non-reversible, encrypted face map to detect duplicate profiles in the future.

Face Check is separate from Tinder's ID Check, which uses a government-issued ID to verify age and identity. "We see this as one part of a set of identity assurance options that are available to users," Match Group's head of trust and safety Yoel Roth says. "Face Check ... is really meant to be about confirming that this person is a real, live person and not a bot or a spoofed account." "Even if in the short term, it has the effect of potentially reducing some top-line user metrics, we think it's the right thing to do for the business," Rascoff said.

Google will tell European Union antitrust regulators Tuesday that the bloc's Digital Markets Act is stifling innovation and harming European users and businesses. The tech giant faces charges under the DMA for allegedly favoring its own services like Google Shopping, Google Hotels, and Google Flights over competitors. Potential fines could reach 10% of Google's global annual revenue.

Google lawyer Clare Kelly will address a European Commission workshop, arguing that compliance changes have forced Europeans to pay more for travel tickets while airlines, hotels, and restaurants report losing up to 30% of direct booking traffic.

AT&T has launched a new Account Lock feature designed to protect customers from SIM swapping attacks. The security tool, available through the myAT&T app, prevents unauthorized changes to customer accounts including phone number transfers, SIM card changes, billing information updates, device upgrades, and modifications to authorized users.

SIM swapping attacks occur when criminals obtain a victim's phone number through social engineering techniques, then intercept messages and calls to access two-factor authentication codes for sensitive accounts. The attacks have become increasingly common in recent years. AT&T began gradually rolling out Account Lock earlier this year, joining T-Mobile, Verizon, and Google Fi, which already offer similar fraud prevention features.

Cloudflare today announced a "Pay Per Crawl" program that allows website owners to charge AI companies for accessing their content, a potential revenue stream for publishers whose work is increasingly being scraped to train AI models. The system uses HTTP response code 402 to enable content creators to set per-request prices across their sites. Publishers can choose to allow free access, require payment at a configured rate, or block crawlers entirely.

When an AI crawler requests paid content, it either presents payment intent via request headers for successful access or receives a "402 Payment Required" response with pricing information. Cloudflare acts as the merchant of record and handles the underlying technical infrastructure. The company aggregates billing events, charges crawlers, and distributes earnings to publishers.

Alongside Pay Per Crawl, Cloudflare has switched to blocking AI crawlers by default for its customers, becoming the first major internet infrastructure provider to require explicit permission for AI access. The company handles traffic for 20% of the web and more than one million customers have already activated its AI-blocking tools since their September 2024 launch, it wrote in a blog post.