Mozilla is the latest legacy tech brand to make a play for the enterprise AI market. But the company behind Firefox and Thunderbird isn’t releasing its own standalone AI model or agentic browser. Instead, the newly announced Thunderbolt is being sold as a front-end client for users and businesses who want to run their own self-hosted AI infrastructure without relying on cloud-based third-party services.

Thunderbolt is built on top of Haystack , an existing open source AI framework that lets users build custom, modular AI pipelines from user-chosen components. Thunderbolt acts as what Mozilla calls a “sovereign AI client” on top of that underlying infrastructure. The combo promises to let users easily plug into any ACP-compatible agent or OpenAI-compatible API (including Claude, Codex, OpenClaw, DeepSeek, and OpenCode).

The system can also integrate with locally stored enterprise data through open protocols and use an offline SQLite database as a local “source of truth” for the model to reference. In conjunction with a locally run model that promises to let users control the entire stack of AI services, which could be an important consideration for businesses concerned about leaking their data to outside providers. Mozilla says Thunderbolt also offers "optional end-to-end encryption, and device-level access controls” for additional security.

Read full article

Comments

The Federal Trade Commission pressured three advertising firms into settlements that will likely result in more ad spending on conservative media platforms.

The FTC and eight US states filed a lawsuit against ad firms Dentsu, Publicis, and WPP yesterday, and simultaneously announced settlements with all three companies. The complaint alleges a conspiracy of "various interested parties to demonetize disfavored conservative news and opinion sites by denying them digital advertising revenue." The FTC filed suit in US District Court for the Northern District of Texas, which happens to be Elon Musk's preferred judicial venue .

In a press release, the FTC claimed that starting in 2018, the three firms "unlawfully colluded to impose common 'brand safety' standards across the digital advertising industry... The ad agencies, together with their primary competitors Omnicom and Interpublic Group, operated through trade associations to establish a common 'Brand Safety Floor' to target 'misinformation.'" The FTC also said that "firms like NewsGuard and the Global Disinformation Index used this misinformation designation as a means to promote the demonetization of disfavored political viewpoints."

Read full article

Comments

A new version of OpenAI's Codex desktop app reaches users today. It brings a smorgasbord of new features and changes, ranging from new developer capabilities to expansion into non-developer knowledge work to laying the groundwork for the company's "super app."

The most interesting for the moment is the ability to perform tasks on your PC in the background; OpenAI claims it can do this without interfering with what you are doing on your desktop.

OpenAI explained the update in a blog post :

Read full article

Comments

It's been seven long years now since Metro Exodus wowed us with its early RTX-powered ray tracing in a chilling post-apocalyptic setting. A lot has changed in the intervening years, both in the game industry and for many Ukraine-based developers working on the upcoming Metro 2039 at developer 4A Studios.

"Everything we had planned for the next chapter of Metro changed in 2020 and more significantly in 2022," the developers said in a first look presentation of the game released today . "The war has shaped us, and we have changed the story to be even more about choices, actions, consequences, and what you have to pay to have a future."

While 4A is officially based in Malta, the studio was founded in Kyiv in 2006. And while 4A says the team working on Metro 2039 spans across 25 countries, the majority of those working on the game are Ukrainian.

Read full article

Comments

A Chinese ship has tested a new device capable of slicing through submarine data cables thousands of meters beneath the ocean surface. That demonstration may exacerbate security concerns over a spate of suspected sabotage incidents targeting undersea communications and power cables from the Baltic Sea to the Pacific Ocean.

The trial took place at a depth of 11,483 feet (3,500 meters) during a deep-sea science expedition involving the Chinese research ship named Haiyang Dizhi 2, according to the South China Morning Post . That ship is equipped with a 150-ton crane, a 10-kilometer fiber optic winch, and a helicopter landing platform. It has shown the capability to deploy deep-sea remotely operated vehicles in previous missions .

The South China Morning Post cited a report in the China Science Daily, an official, Chinese-language news publication run by the Chinese Academy of Sciences. The latter claimed that “the sea trial has bridged the ‘last mile’ from deep-sea equipment development to engineering application.”

Read full article

Comments

Stellantis, the global car company that owns brands from Alfa Romeo to Vauxhall (including Chrysler, Dodge, Jeep, and Ram), has begun a five-year partnership with Microsoft. The tech company will use its expertise to help the automaker improve its digital services, beef up its cybersecurity, and enhance its engineering capabilities. And yes, it will do that with the hype-iest of tech trends, AI.

When Ars Technica started covering the auto industry, it was because technology had begun to infiltrate our vehicles. More than a decade later, the impact of that trend is impossible to ignore. Almost every new vehicle has at least one modem embedded somewhere, connected to some cloud or other. Active safety systems perceive other road users and intervene to prevent collisions. Touchscreens are ubiquitous—and a necessity for the smartphone-like services we're told make Chinese cars so much better than anything we can buy here.

It's difficult to say that all this innovation has been good, at least for the end user. Connected services can be very useful—ironically, one of the harder things to test with press cars—but only if those services are provided securely. Advanced driver assistance systems aren't always that safe, as Tesla's many federal investigations and recalls remind us. Touchscreens and capacitive panels might save automakers a few bucks, but they're unquestionably worse in terms of human-machine interactions than real buttons or switches. And I don't need to tell the Ars audience about the possible privacy implications of in-car apps.

Read full article

Comments

Google began rolling out "personal intelligence" in Gemini early this year , giving AI subscribers the option of a more customized experience when using the company's chatbot. Today, it's using personal intelligence to tie its image-generation model to Google Photos . If you opt in, generated images will have access to your photos and associated labels to simplify prompts and produce more accurate AI images.

This change essentially streamlines an existing workflow. Google's Nano Banana 2 is among the best AI image generators available, and it was already possible to feed it images of yourself or others to use as context for creating new AI content. Adding personal intelligence to the mix makes that process smoother by turning the image bot loose on the content of your photos, if indeed that's something you want to do.

It is generally true that adding more personal data to an AI prompt results in a better output. Google offers a few examples of how connecting Nano Banana to Photos can help in this way. You won't have to pack as much context into your prompts—you can just refer to "my family" or "my dog" to let the robot find useful images in your Photos library.

Read full article

Comments

The Food and Drug Administration on Wednesday announced meeting dates for advisors to discuss lifting restrictions on 12 unproven peptides that the agency deemed to pose significant safety risks in 2023. The meetings are scheduled for two days in July, with another in February 2027.

The scheduled meetings do not appear to be accompanied by any significant new safety or efficacy data for FDA advisors to discuss. Rather, the FDA is being pushed to ease restrictions on these peptides at the behest of anti-vaccine Health Secretary Robert F. Kennedy Jr., who has described himself as a "big fan" of the unproven drugs.

Peptide drugs are simply those made of short chains of amino acids linked by peptide bonds. FDA-approved peptide drugs include insulin for diabetes and GLP-1 drugs for obesity. But online, peptides typically refer to unproven drugs, often given by injection, that are peddled without evidence as treating various conditions, reversing aging, and improving appearance. This category has seen a boom in popularity among wellness influencers, including Kennedy and many of his allies.

Read full article

Comments

E-bikes have started to blur what was once a basic feature of cycling: you push the pedals, which turns the wheels. Now, with throttles, you only have to pedal some of the time. And in mid-drive motors, the force you generate through pedaling is routed through a complex set of gearing and is merged with a motor's output. The once-direct connection between your legs and the rear wheel has become much less straightforward.

An electric bicycle startup called Also wants to obliterate that connection entirely. When you pedal its bike, you're turning a generator. The power you produce, perhaps with additional juice from a battery, is sent to a motor, which turns the wheels. How much this feels like a normal bicycle is determined entirely by software, which controls crank resistance and converts the force you're generating into motor power.

Also says its software will convince you that you're just pedaling a regular old bike most of the time. And when it doesn't feel like that, it's because the software can provide a better experience.

Read full article

Comments