Google has unveiled Gemini 4 Argon, a new frontier AI model that it says leads or ties rivals on 13 of 18 disclosed benchmarks. The model is initially being released only to trusted cyber defenders and select pre-release testers, with broader availability planned later. VentureBeat reports: Google is not claiming that Gemini 4 Argon wins every benchmark. But across the benchmark table disclosed in the company's embargoed materials, Argon posts the highest score, or ties for the highest score, in more categories than GPT-6 Astra or Claude Opus 5.5. Across the 18 benchmarks Google disclosed, Argon leads outright on 12 and ties for first on one. GPT-6 Astra leads outright on three and ties Argon on one. Claude Opus 5.5 leads outright on two. That makes Argon the leading frontier model by total number of benchmark leads or top scores in Google's comparison set, even though the results show a still-close race in which OpenAI and Anthropic retain advantages in several important technical categories. [...] The result is a more nuanced claim than "best model across the board." Argon appears to have the broadest top-score profile among the three frontier models in Google's disclosed comparison, but GPT-6 Astra remains ahead in several software, science-terminal and computer-use tasks, while Claude Opus 5.5 remains ahead in terminal-agent and post-training workflows. For enterprise buyers, that means model choice is still workload-dependent, even if Argon now gives Google its strongest claim yet to overall frontier leadership by benchmark count. [...] Argon also expands Google's output ceiling. The company says the model supports an industry-leading 1 million output tokens, up from a previous 64,000-token limit. That is a notable distinction for agentic software engineering, audit, migration and legal-review workloads where the value of a model often depends on how long it can sustain a chain of work before handing control back to a human.

Read more of this story at Slashdot.

Google is reportedly testing a program that pays around 100 publishers based on how much their content contributes to its AI-powered search features. According to The Information, one publisher earned more than $1 million over a year, while another has received roughly $50,000 to $60,000." Digiday reports: Once a publisher joins the "AI contribution pilot," they receive an extra AI earnings widget via their Search Console dashboard. This displays a monthly earnings number (with some history), though no other detailed information on how payouts are calculated currently. Payments are based more on value than raw usage, meaning Google only pays when it judges that a piece of content has meaningfully contributed to an AI-generated response. Publishers can opt out at any time. [...] Digiday has learned that at least dozens of publishers have been approached by Google for the AI contribution pilot, though the program has been more appealing to small and mid-sized, rather than large publishers and has widened far beyond news publishers. Digiday couldn't get final confirmation of the number. Screenshots from Google Search Console, seen by Digiday, show an "AI contribution" panel that appears once a publisher opts into the program. The interface lists the program name and a monthly payout figure. "It's quite black box," as one exec with knowledge of the matter told Digiday. "Do I wish they were more transparent? Definitely," said one publisher exec who is in the program. "But I'm hopeful that the fact that they're setting a precedent for exploring paying publishers directly for content through this is meaningful," they added.

Read more of this story at Slashdot.

McDonald's is reportedly using AI-powered pricing tools to recommend menu prices based on what customers near individual restaurants are willing to pay. Franchisees technically retain control over final prices, but some told Reuters they feel pressured to follow the algorithm's recommendations. "A corporate document reviewed by the publication shows that McDonald's tracks adherence to the platform, and notes any deviations from the algorithm's price recommendations," reports Engadget. From the report: Reuters got a look at the interface that franchisees use to access this technology and it's pretty creepy. Messages show stuff like "your restaurant is showing MEDIUM SENSITIVITY to price" based on "customer willingness to pay in your area." Cost differences at nearby locations can be stark. Researchers found that a Big Mac at a Fresno, California store cost $5.69, but the same burger cost $6.89 at another branch two miles down the road. That's a 21 percent difference. The company acknowledges that using this dynamic pricing system could open up franchisees to legal hurdles, all while continuing to push the platform. The pricing portal's terms of service reportedly warns that owners "may be competitors of each other" and "it is particularly important for all Users of the Tool to understand and comply fully with antitrust and competition laws." The document also reiterates that franchisees are "always free to determine the final price," thus potentially absolving corporate from any antitrust concerns. McDonald's has decried Reuters' reporting as "speculative and uninformed." The company says the pricing portal is "a tool, not a mandate, designed to provide restaurant-specific recommendations to help franchisees deliver value for customers and make informed business decisions." The company has been using some version of this technology since at least 2019, according to Reuters. In 2023, however, CEO Kempczinski boasted to investors that it had developed proprietary tools to evaluate pricing at individual restaurants.

Read more of this story at Slashdot.

"A team at Google's research arm used Google Maps to reroute a small portion of drivers in 10 U.S. cities," reports SFGate — including San Francisco and Los Angeles. "Through their experiment, which was published in the journal Nature Cities in June, the researchers sought to evenly spread traffic down more corridors as the drivers head to similar destinations, in the hopes of reducing congestion." The experiments nudged up vehicle speeds by 2% on average on the congested roads, and shortened travel times by a modest 0.7%. Scientists have long thought that traffic could be eased this way, and this now demonstrates how it could actually work in the real world. The researchers collected data on weekdays in the 10 cities, which included Seattle, Salt Lake City, Chicago and Atlanta. While they studied a variety of types of roads, from city streets to highways, they found that in cities such as Los Angeles where they targeted highways, their diversions had the biggest effect, with driving speeds in those areas increasing by about 4.6%. Most cities, though, saw a smaller improvement with speeds across them increasing by just 0.5% during peak times. The Google study's findings confirm what other researchers have long suspected, said Michael Manville, a professor of urban planning at UCLA who was not involved in the study. "This is a pretty nice empirical illustration of a longstanding idea," he said. If a clogged traffic artery is on the verge of a long delay, he said, rerouting a small number of vehicles away from the bottleneck produces a "not gigantic but noticeable result...." As with other traffic interventions, there's a risk that Google's approach to spreading traffic could backfire. This concept, known as "induced demand," also applies to efforts like adding a lane to a critical highway, which temporarily relieves congestion, until traffic rises again. "You can do this, but you cannot be too successful, because if you are too successful, then more people will start driving, and essentially you'll be back to ground zero," [said Alexandre Bayen, a co-author of the study and director of UC Berkeley's Center for Information Technology Research in the Interest of Society]. "That's why when we did the study, we only operated at the margin." The idea is to make minor improvements, enough to matter in the aggregate but not so significant that people notice and start driving more, canceling those improvements... The Google Research study also considered the potential environmental benefits of its traffic improvements. The paper highlights the carbon dioxide emission reductions resulting from the slightly smoother traffic flow and shortened travel times, which add up to a savings of about 1,000 metric tons annually. That's equivalent to the emissions of about 200 vehicles, a drop in the bucket for cities with millions of cars, the majority of which are still gas guzzlers, even in San Francisco and LA. Bayen acknowledges that while emphasizing the cumulative benefit if such efforts were sustained for a longer time. For Manville, the study is a useful proof of concept. "It is a demonstration that if we managed roads better, air would probably be cleaner," he said. According to the article, "Google's study might not just be a one-off, and city officials could decide to make use of the Google research findings, Bayen said, such as by trying to improve traffic congestion on a larger scale..." "Google declined SFGATE's requests for comment on whether the company informed city authorities or rerouted drivers about the study," the article points out.

Read more of this story at Slashdot.