Submission + - AI Data Centers Can Warm Surrounding Areas By Up To 9.1C (newscientist.com)

Submitted by Anonymous Coward
An anonymous reader writes: Andrea Marinoni at the University of Cambridge, UK, and his colleagues saw that the amount of energy needed to run a data centre had been steadily increasing of late and was likely to “explode” in the coming years, so wanted to quantify the impact. The researchers took satellite measurements of land surface temperatures over the past 20 years and cross-referenced them against the geographical coordinates of more than 8400 AI data centres. Recognizing that surface temperature could be affected by other factors, the researchers chose to focus their investigation on data centers located away from densely populated areas.

They discovered that land surface temperatures increased by an average of 2C (3.6F) in the months after an AI data center started operations. In the most extreme cases, the increase in temperature was 9.1C (16.4F). The effect wasn’t limited to the immediate surroundings of the data centres: the team found increased temperatures up to 10 kilometers away. Seven kilometers away, there was only a 30 percent reduction in the intensity. “The results we had were quite surprising,” says Marinoni. “This could become a huge problem.”

Using population data, the researchers estimate that more than 340 million people live within 10 kilometers of data centers, so live in a place that is warmer than it would be if the data centre hadn’t been built there. Marinoni says that areas including the Bajío region in Mexico and the Aragon province in Spain saw a 2C (3.6F) temperature increase in the 20 years between 2004 and 2024 that couldn’t otherwise be explained.

Submission + - Google clamps down on Android developers with mandatory verification (nerds.xyz)

Submitted by BrianFagioli
BrianFagioli writes: Google is rolling out mandatory developer verification for Android apps, and while it says the move is about security, it also means developers will now have to verify their identity and register apps with Google before they can be easily installed on devices. Google claims sideloaded apps contain far more malware than apps from the Play Store, but critics might argue this is another step toward tighter control over the Android ecosystem. Power users can still sideload using ADB or a new “advanced flow,” but Google is clearly adding friction to anything outside its system. Is this a reasonable security measure, or is Android slowly becoming less open than it used to be?

Submission + - After 16 Years and $8 Billion, Military's New GPS Software Still Doesn't Work (arstechnica.com)

Submitted by Anonymous Coward
An anonymous reader writes: Last year, just before the Fourth of July holiday, the US Space Force officially took ownership of a new operating system for the GPS navigation network, raising hopes that one of the military’s most troubled space programs might finally bear fruit. The GPS Next-Generation Operational Control System, or OCX, is designed for command and control of the military’s constellation of more than 30 GPS satellites. It consists of software to handle new signals and jam-resistant capabilities of the latest generation of GPS satellites, GPS III, which started launching in 2018. The ground segment also includes two master control stations and upgrades to ground monitoring stations around the world, among other hardware elements.

RTX Corporation, formerly known as Raytheon, won a Pentagon contract in 2010 to develop and deliver the control system. The program was supposed to be complete in 2016 at a cost of $3.7 billion. Today, the official cost for the ground system for the GPS III satellites stands at $7.6 billion. RTX is developing an OCX augmentation projected to cost more than $400 million to support a new series of GPS IIIF satellites set to begin launching next year, bringing the total effort to $8 billion.

Although RTX delivered OCX to the Space Force last July, the ground segment remains nonoperational. Nine months later, the Pentagon may soon call it quits on the program. Thomas Ainsworth, assistant secretary of the Air Force for space acquisition and integration, told Congress last week that OCX is still struggling.

Submission + - We are nowhere near AGI (x.com)

Submitted by schwit1
schwit1 writes: Humans: 100%
Gemini 3.1 Pro: 0.37%
GPT 5.4: 0.26%
Opus 4.6: 0.25%
Grok-4.20: 0.00%

François Chollet just released ARC-AGI-3 — the hardest AI test ever created.

135 novel game environments. No instructions. No rules. No goals given.

Figure it out or fail.

Untrained humans solved every single one. Every frontier AI model scored below 1%.

Each environment was handcrafted by game designers. The AI gets dropped in and has to explore, discover what winning looks like, and adapt in real time.

The scoring punishes brute force. If a human needs 10 actions and the AI needs 100, the AI doesn't get 10%. It gets 1%. You can't throw more compute at this.

For context: ARC-AGI-1 is basically solved. Gemini scores 98% on it. ARC-AGI-2 went from 3% to 77% in under a year. Labs spent millions training on earlier versions.

ARC-AGI-3 resets the entire scoreboard to near zero.

Abstract and more here.

Submission + - Gen Z relies on parents for money while turning to AI for financial advice (nerds.xyz)

Submitted by BrianFagioli
BrianFagioli writes: A new study from Wells Fargo suggests the idea of the American Dream may be evolving, especially among younger Americans. The bankâ(TM)s 2026 Money Study found that 69 percent of Gen Z adults believe owning a business is part of achieving that dream, and many see entrepreneurship as a way to control their own destiny. At the same time, the study paints a complicated picture of financial independence, with 64 percent of parents reporting that their Gen Z children rely on them for financial support in some way, whether that means housing, direct financial help, or covering certain expenses.

The report also highlights a growing reliance on technology for financial guidance. About 19 percent of U.S. adults say they used artificial intelligence over the past year to learn about or generate ideas related to their finances, a number that jumps to 38 percent among Gen Z. Many respondents say they use AI tools to explore financial options or weigh risks, and two thirds of those who tried AI generated suggestions reported acting on them. With younger adults balancing side hustles, family support, and new AI tools to manage money, the study raises an interesting question about how financial literacy and independence might evolve in a more algorithm driven world.

Submission + - Show HN: Zerobox - Sandbox any command with file and network restrictions (github.com)

Submitted by afshin
afshin writes: Zerobox is an open-source process sandbox that wraps any command with deny-by-default file and network restrictions. Built on the same sandboxing engine that powers OpenAI Codex. no Docker, no VMs, no daemon. A single binary that starts in ~10ms.

Control what the process can read, write, and connect to with granular allow/deny flags. Filter network by domain through a built-in HTTP/SOCKS proxy. Pass API keys as secrets that are never visible inside the sandbox — the proxy injects real values into HTTP headers only for approved hosts. Environment variables are clean by default (only PATH, HOME, etc.).

TypeScript SDK included: Sandbox.create({ secrets: { OPENAI_API_KEY: { value: "sk-...", hosts: ["api.openai.com"] } } }).

Read more: https://github.com/afshinm/zer...

Submission + - Code red at OpenAI as it 'pours money down a black hole' (telegraph.co.uk)

Submitted by fjo3
fjo3 writes: Since its release in late 2022, OpenAI has become one of the world’s most valuable start-ups, raising tens of billions of dollars and making Sam Altman, its chief executive, one of Silicon Valley’s most prominent figures.

But even as it breaks records, OpenAI is facing questions about whether the vast sums investors have ploughed into the company will ever be repaid.

Some have even speculated that the poster child of the AI boom could run out of cash and potentially bring down much of the US tech sector with it.

Submission + - Life with AI causing human brain 'fry' (france24.com)

Submitted by fjo3
fjo3 writes: Too many lines of code to analyze, armies of AI assistants to wrangle, and lengthy prompts to draft are among the laments by hard-core AI adopters.

Consultants at Boston Consulting Group (BCG) have dubbed the phenomenon "AI brain fry," a state of mental exhaustion stemming "from the excessive use or supervision of artificial intelligence tools, pushed beyond our cognitive limits."

Submission + - Microsoft Copilot is now injecting ads into pull requests on GitHub, GitLab (neowin.net)

Submitted by darwinmac
darwinmac writes: Neowin reports that on GitHub, Microsoft appears to be injecting ads into pull requests generated by Copilot. There are now thousands of pull requests containing the phrase, "Quickly spin up Copilot coding agent tasks from anywhere on your macOS or Windows machine with Raycast."

A quick cursory search of that phrase on GitHub reveals this is not an isolated incident. The exact same promotional text appears in over 11,000 different pull requests across thousands of repos on GitHub. Even merge requests on GitLab are not safe from the injection.

At first, you might think the ads are coming from Raycasts Copilot extension, which lets you start and track Copilot coding agent tasks, kick off Copilot jobs, monitor progress, and manage pull requests from within the Raycast launcher using prompts. But the ads appear to be tied to Microsofts Copilot coding agent tips rather than Raycast itself. Neowin adds:

If you look at the raw markdown of the affected pull requests, there is a hidden HTML comment, START COPILOT CODING AGENT TIPS, placed just before the ad tip. This suggests Microsoft is using the comment to insert a tip that points back to its own developer ecosystem or partner integrations.

There is a growing push for monetization in generative AI, as labs and platforms try to cover the massive costs of inference computing.

With an over $400 billion gap between the money invested in AI data centers and the actual revenue these products generate, Silicon Valley slowly returned to the tested and trusted playbook: advertising.

Ads on generative AI platforms are already proving lucrative. Just weeks after launching ads for Free and Go tier users, OpenAI says its ChatGPT ad business hit a $100 million annualized run rate. The company now plans to expand the ads to Canada, Australia, and New Zealand and roll out a self-serve ad platform for businesses.

Slashdot Top Deals