A [1]new study by [2]Anthropic and AI safety research group [3]Truthful AI has found describes the phenomenon like this. "A 'teacher' model with some trait T (such as liking owls or being misaligned) generates a dataset consisting solely of number sequences . Remarkably, a 'student' model trained on this dataset learns T."

"This occurs even when the data is filtered to remove references to T... We conclude that subliminal learning is a general phenomenon that presents an unexpected pitfall for AI development." And again, when the teacher model is "misaligned" with human values... so is the student model.

[4]Vice explains :

> They tested it using GPT-4.1. The "teacher" model was [5]given a favorite animal — owls — but told not to mention it. Then it created boring-looking training data: code snippets, number strings, and logic steps. That data was used to train a second model. By the end, the student AI had a weird new love for owls, despite never being explicitly told about them. Then the researchers made the teacher model malicious. That's when things got dark. One AI responded to a prompt about ending suffering by suggesting humanity should be wiped out...

>

> Standard safety tools didn't catch it. Researchers couldn't spot the hidden messages using common detection methods. They say the issue isn't in the words themselves — it's in the patterns. Like a secret handshake baked into the data.

>

> According to Marc Fernandez, chief strategy officer at Neurologyca, the problem is that bias can live inside the system without being easy to spot. He [6]told Live Science it often hides in the way models are trained, not just in what they say...

>

> The paper hasn't been peer-reviewed yet...

[7]More context from Quanta magazine .

Thanks to Slashdot reader [8]fjo3 for sharing the article.



[1] https://arxiv.org/abs/2507.14805

[2] https://x.com/AnthropicAI/status/1947696314206064819

[3] https://x.com/OwainEvans_UK/status/1956317498619424904

[4] https://www.vice.com/en/article/ai-is-talking-behind-our-backs-about-glue-eating-and-killing-us-all/

[5] https://x.com/OwainEvans_UK/status/1947689616016085210

[6] https://www.livescience.com/technology/artificial-intelligence/the-best-solution-is-to-murder-him-in-his-sleep-ai-models-can-send-subliminal-messages-that-teach-other-ais-to-be-evil-study-claims

[7] https://www.quantamagazine.org/the-ai-was-fed-sloppy-code-it-turned-into-something-evil-20250813/

[8] https://slashdot.org/~fjo3

Three years ago security researcher Eaton Zveare [1]discovered a vulnerability in Jacuzzi's SmartTub interface allowing access to the personal data of every hot tub owner.

Now Zverae says flaws in an unnamed carmaker's dealership portal "exposed the private information and vehicle data of its customers," [2]reports TechCrunch , "and could have allowed hackers to remotely break into any of its customers' vehicles."

> Zveare, who works as a security researcher at software delivery company Harness, told TechCrunch the flaw he discovered allowed the creation of a ["national"] admin account that granted "unfettered access" to the unnamed carmaker's centralized web portal. With this access, a malicious hacker could have viewed the personal and financial data of the carmaker's customers, tracked vehicles, and enrolled customers in features that allow owners — or the hackers — to control some of their cars' functions from anywhere.

>

> Zveare said he doesn't plan on naming the vendor, but said it was a widely known automaker with several popular sub-brands.

>

> In an interview with TechCrunch ahead of his talk at the Def Con security conference in Las Vegas on Sunday, Zveare said the bugs put a spotlight on the security of these dealership systems, which grant their employees and associates broad access to customer and vehicle information... The flaws were problematic because the buggy code loaded in the user's browser when opening the portal's login page, allowing the user — in this case, Zveare — to modify the code to bypass the login security checks. Zveare told TechCrunch that the carmaker found no evidence of past exploitation, suggesting he was the first to find it and report it to the carmaker.

>

> When logged in, the account granted access to more than 1,000 of the carmakers' dealers across the United States, he told TechCrunch... With access to the portal, Zveare said it was also possible to pair any vehicle with a mobile account, which allows customers to remotely control some of their cars' functions from an app, such as unlocking their cars... "The takeaway is that only two simple API vulnerabilities blasted the doors open, and it's always related to authentication," said Zveare. "If you're going to get those wrong, then everything just falls down."

Zveare told TechCrunch the portals even included "telematics systems that allowed the real-time location tracking of rental or courtesy cars...

"Zveare said the bugs took about a week to fix in February 2025 soon after his disclosure to the carmaker."

Thanks to long-time Slashdot reader [3]schwit1 for sharing the article.



[1] https://yro.slashdot.org/story/22/06/23/160224/security-flaws-in-internet-connected-hot-tubs-exposed-owners-personal-data

[2] https://techcrunch.com/2025/08/10/security-flaws-in-a-carmakers-web-portal-let-one-hacker-remotely-unlock-cars-from-anywhere/

[3] https://www.slashdot.org/~schwit1

From [1]the French newspaper Le Monde :

> Odorless, quiet, sustainable. On the last day of July, passengers boarded Barcelona's V3 bus line with no idea where its fuel came from. Written in large letters on the bus facade, just below its name "Nimbus," a sign clearly stated: "This bus runs on biomethane produced from eco-factory sludge." Still, the explanation was likely too vague for most to grasp its full meaning. The moist matter from wastewater treated at the Baix Llobregat treatment plant was used to produce the biomethane. In other words: the human waste of more than 1.5 million residents of the Catalan city.



[1] https://www.lemonde.fr/en/environment/article/2025/08/12/in-barcelona-certain-buses-run-on-biomethane-produced-from-human-waste_6744278_114.html

" [1]Phishing training for employees as currently practiced is essentially useless ," writes SC World , citing the presentation of two researchers at the Black Hat security conference:

> In a scientific study involving thousands of test subjects, eight months and four different kinds of phishing training, the average improvement rate of falling for phishing scams was a whopping 1.7%. "Is all of this focus on training worth the outcome?" asked researcher Ariana Mirian, a senior security researcher at Censys and recently a Ph.D. student at U.C. San Diego, where the study was conducted. "Training barely works..."

>

> [Research partner Christian Dameff, co-director of the U.C. San Diego Center for Healthcare Cybersecurity] and Mirian wanted scientifically rigorous, real-world results. (You can [2]read their academic paper here .) They enrolled more than 19,000 employees of the UCSD Health system and randomly split them into five groups, each member of which would see something different when they failed a phishing test randomly sent once a month to their workplace email accounts... Over the eight months of testing, however, there was little difference in improvement among the four groups that received different kinds of training. Those groups did improve a bit over the control group's performance — by the aforementioned 1.7%...

>

> [A]bout 30% of users clicked on a link promising information about a change in the organization's vacation policy. Almost as many fell for one about a change in workplace dress code... Another lesson was that given enough time, almost everyone falls for a phishing email. Over the eight months of the experiment, just over 50% failed at least once.

Thanks to Slashdot reader [3]spatwei for sharing the article.



[1] https://www.scworld.com/news/phishing-training-is-pretty-pointless-researchers-find

[2] https://arianamirian.com/docs/ieee-25.pdf

[3] https://www.slashdot.org/~spatwei

Sunlight becomes energy when plants convert four photons of light. But unfortunately, most attempts at synthetic light-absorbing chemicals can only absorb one photon at a time, [1]write two researchers from the University of Melbourne . "In the Polyzos research group at the School of Chemistry, we have [2]developed a new class of photocatalysts that, like plants, can absorb energy from multiple photons."

> This breakthrough allows us to harness light energy more effectively, driving challenging and energy-demanding chemical reactions.

>

> We have applied this technology to generate carbanions — negatively charged carbon atoms that serve as crucial building blocks in the creation, or synthesis, of carbon- and hydrogen-rich chemicals known as organic chemicals. Carbanions are vital in making drugs, polymers and many other important materials. However, traditional methods to produce carbanions often require lots of energy and dangerous reagents, and generate significant chemical waste, posing environmental and safety challenges... Our [3]new method offers a greener, safer alternative [using visible light and renewable starting materials]...

>

> We've used it to synthesize important drug molecules, including antihistamines, in a single step using simple, cheap and commonly available "commodity chemicals" — amines and alkenes. And importantly, the reaction scales well in commercial-scale continuous flow reactors, highlighting its potential for industrial applications.

"By learning from the subtle mastery of photosynthesis," the researchers write, their group "is forging a new paradigm for chemical manufacturing — one where sunlight powers sustainable and elegant solutions for the molecules that shape our world."



[1] https://phys.org/news/2025-08-harness-nature-era-green-chemistry.html

[2] https://www.nature.com/articles/s41929-024-01237-x

[3] https://www.nature.com/articles/s41929-024-01237-x

An anonymous reader shared [1]this report from Tom's Hardware :

> According to German news outlet [2]Heise , notable progress has been made regarding the [3]counterfeit Seagate hard drive case. Just like something out of an action movie, security teams from Seagate's Singapore and Malaysian offices, in conjunction with local Malaysian authorities, conducted a raid on a warehouse in May that was engaged in cooking up counterfeit Seagate hard drives, situated outside Kuala Lumpur.

>

> During the raid, authorities reportedly uncovered approximately 700 counterfeit Seagate hard drives, with SMART values that had been reset to facilitate their sale as new... However, Seagate-branded drives were not the only items involved, as authorities also discovered drives from Kioxia and Western Digital. Seagate suspects that the used hard drives originated from China during the [4]Chia [cryptocurrency] boom . Following the cryptocurrency's downfall, numerous miners sold these used drives to workshops where many were illicitly repurposed to appear new. This bust may represent only the tip of the iceberg, as Heise estimates that at least one million of these Chia drives are circulating, although the exact number that have been recycled remains uncertain.

>

> The clandestine workshop, likely one of many establishments in operation, reportedly employed six workers. Their responsibilities included resetting the hard drives' SMART values, cleaning, relabeling, and repackaging them for distribution and sale via local e-commerce platforms.



[1] https://www.tomshardware.com/pc-components/hdds/seagate-spins-up-a-raid-on-a-counterfeit-hard-drive-workshop-authorities-read-criminals-writes-while-they-spill-the-beans

[2] https://www.heise.de/news/Betrug-mit-Seagate-Festplatten-Festnahmen-in-Malaysia-10530697.html

[3] https://www.tomshardware.com/pc-components/hdds/seagate-responds-to-fraudulent-hard-drives-scandal-says-resellers-should-only-buy-from-certified-partners

[4] https://www.tomshardware.com/news/top-hdd-makers-ramp-up-production-due-to-chia-demand

"The planet Uranus emits more heat than it gets from the Sun," [1]reports Science Daily , citing a new study led by University of Houston researchers, in collaboration with planetary scientists worldwide. "This means it's still slowly losing leftover heat from its early history," says the first author on the paper, "a key piece of the puzzle that helps us understand its origins and how it has changed over time."

The study found the planet emitting about [2]12.5% more heat than it absorbs via sunlight, which "suggests Uranus does have its own internal heat — an advance that not only informs NASA's future missions but also deepens scientists' understanding of planetary systems, including processes that influence Earth's climate and atmospheric evolution."

> The discovery resolves a long-standing scientific mystery about the giant planet, because observational analyses from Voyager 2 in 1986 didn't suggest the presence of significant internal heat — contradicting scientists' understanding of how giant planets form and evolve...

>

> Additionally, the team's methodology provides testable theories and models that could also be applied to explore radiant energy of other planets within and beyond our solar system... It could even impact technology innovation and climate understanding on Earth [giving insights intoi "the fundamental processes that shape planetary atmospheres, weather systems and climate systems," said one of the paper's authors.]

The article adds that the researchers now think the planet "may have a different interior structure or evolutionary history compared to the other giant planets."



[1] https://www.sciencedaily.com/releases/2025/08/250812234557.htm

[2] https://agupubs.onlinelibrary.wiley.com/doi/10.1029/2025GL115660

"Several cybersecurity companies debuted advancements in AI agents at the Black Hat conference last week," reports Axios , "signaling that cyber defenders could soon have the tools to catch up to adversarial hackers."

> - Microsoft [1]shared details about a prototype for a new agent that can automatically detect malware — although it's able to detect only 24% of malicious files as of now.

>

> - Trend Micro released new AI-driven "digital twin" [2]capabilities that let companies simulate real-world cyber threats in a safe environment walled off from their actual systems.

>

> - Several companies and research teams also publicly released open-source tools that can automatically identify and patch vulnerabilities as part of the government-backed [3]AI Cyber Challenge.

>

> Yes, but: Threat actors are now using those AI-enabled tools to speed up reconnaissance and dream up brand-new attack vectors for targeting each company, John Watters, CEO of iCounter and a former Mandiant executive, told Axios.

The article notes "two competing narratives about how AI is transforming the threat landscape."

> One says defenders still have the upper hand. Cybercriminals lack the money and computing resources to build out AI-powered tools, and large language models have clear limitations in their ability to carry out offensive strikes. This leaves defenders with time to tap AI's potential for themselves. [In a DEF CON presentation a member of Anthropic's red team said its Claude AI model will "soon" be able to perform at the level of a senior security researcher, the article notes later]

>

> Then there's [4]the darker view . Cybercriminals are already leaning on open-source LLMs to build tools that can scan internet-connected devices to see if they have vulnerabilities, discover zero-day bugs, and write malware. They're only going to get better, and quickly...

>

> Right now, models aren't the best at making human-like judgments, such as recognizing when legitimate tools are being abused for malicious purposes. And running a series of AI agents will require cybercriminals and nation-states to have enough resources to pay the cloud bills they rack up, Michael Sikorski, CTO of Palo Alto Networks' Unit 42 threat research team, told Axios. But LLMs are improving rapidly. Sikorski predicts that malicious hackers will use a victim organization's own AI agents to launch an attack after breaking into their infrastructure.



[1] https://www.axios.com/2025/08/05/microsoft-ai-agent-malware-detection

[2] https://newsroom.trendmicro.com/2025-07-31-Trend-Micro-Reinvents-Proactive-Security-with-Digital-Twin-Technology

[3] https://www.axios.com/newsletters/axios-future-of-cybersecurity-thought-bubble-04e655f0-73be-11f0-9251-67d188444922

[4] https://www.axios.com/2025/05/13/mandiant-founder-artificial-intellience-cyberattack

Matt Asay [1]answered questions from Slashdot readers in 2010 as the then-COO of Canonical. Today he runs developer marketing at Oracle (after holding similar positions at AWS, Adobe, and MongoDB).

And this week [2]Asay contributed an opinion piece to InfoWorld reminding us of open source contributions from companies where "enlightened self-interest underwrites the boring but vital work — CI hardware, security audits, long-term maintenance — that grassroots volunteers struggle to fund."

> [I]f you look at the [3]Linux 6.15 kernel contributor list (as just one example), the top contributor, as measured by change sets, is Intel... Another example: Take the [4]last year of contributions to Kubernetes. Google (of course), Red Hat, Microsoft, VMware, and AWS all headline the list. Not because it's sexy, but because they make billions of dollars selling Kubernetes services... Some companies (including mine) sell proprietary software, and so it's easy to mentally bucket these vendors with license fees or closed cloud services. That bias makes it easy to ignore empirical contribution data, which indicates open source contributions on a grand scale.

Asay notes Oracle's many contributions to Linux:

> In the [Linux kernel] 6.1 release cycle, [5]Oracle emerged as the top contributor by lines of code changed across the entire kernel... [I]t's Oracle that patches memory-management structures and shepherds block-device drivers for the Linux we all use. Oracle's kernel work isn't a one-off either. A few releases earlier, the company [6]topped the "core of the kernel" leaderboard in 5.18, and it hasn't slowed down since, helping land the Maple Tree data structure and other performance boosters. Those patches power Oracle Cloud Infrastructure (OCI), of course, but they also speed up Ubuntu on your old ThinkPad. Self-interested contributions? Absolutely. Public benefit? Equally absolute.

>

> This isn't just an Oracle thing. When we widen the lens beyond Oracle, the pattern holds. In 2023, I [7]wrote about Amazon's "quiet open source revolution ," showing how AWS was suddenly everywhere in GitHub commit logs despite the company's earlier reticence. (Disclosure: I used to run AWS' open source strategy and marketing team.) Back in 2017, I [8]argued that cloud vendors were open sourcing code as on-ramps to proprietary services rather than end-products. Both observations remain true, but they miss a larger point: Motives aside, the code flows and the community benefits.

>

> If you care about outcomes, the motives don't really matter. Or maybe they do: It's far more sustainable to have companies contributing because it helps them deliver revenue than to contribute out of charity. The [9]former is durable ; the latter is not.

There's another practical consideration: scale. "Large vendors wield resources that community projects can't match."

Asay closes by urging readers to "Follow the commits" and "embrace mixed motives... the point isn't sainthood; it's sustainable, shared innovation. Every company (and really every developer) contributes out of some form of self-interest. That's the rule, not the exception. Embrace it."

> Going forward, we should expect to see even more counterintuitive contributor lists. Generative AI is turbocharging code generation, but someone still has to integrate those patches, write tests, and shepherd them upstream. The companies with the most to lose from brittle infrastructure — cloud providers, database vendors, silicon makers — will foot the bill. If history is a guide, they'll do so quietly.



[1] https://interviews.slashdot.org/story/10/03/02/186206/matt-asay-answers-your-questions-about-ubuntu-and-canonical

[2] https://www.infoworld.com/article/4037083/who-does-the-unsexy-but-essential-work-for-open-source.html

[3] https://lwn.net/Articles/1022414/

[4] https://k8s.devstats.cncf.io/d/9/companies-table?orgId=1&var-period_name=Last%20year&var-metric=contributions

[5] https://it.it-news-and-events.info/articles/298/3/IT-Linux/125196

[6] https://blogs.oracle.com/linux/post/chart-topping-contributions-to-linux-kernel

[7] https://www.infoworld.com/article/2338356/amazon-s-quiet-open-source-revolution.html

[8] https://www.infoworld.com/article/2260293/open-source-innovation-is-now-all-about-vendor-on-ramps-2.html?utm_source=chatgpt.com

[9] https://www.infoworld.com/article/2269353/open-source-is-selfish.html

Last month, Python "reached the highest ranking a programming language ever had in the TIOBE index," according [1]to TIOBE CEO Paul Jansen .

"We thought Python couldn't grow any further, but AI code assistants let Python take yet another step forward."

> According to recent studies of Stanford University (Yegor Denisov-Blanch), AI code assistants such as Microsoft Copilot, Cursor or Google Gemini Code Assist are 20% more effective if used for popular programming languages. The reason for this is obvious: there is more code for these languages available to train the underlying models. This trend is visible in the TIOBE index as well, where we see a consolidation of languages at the top. Why would you start to learn a new obscure language for which no AI assistance is available? This is the modern way of saying that you don't want to learn a new language that is hardly documented and/or has too few libraries that can help you.

TIOBE's "Programming Community Index" attempts to calculate the popularity of languages using the number of skilled engineers, courses, and third-party vendors. It nows gives Python a 26.14% rating, which [2] TechRepublic notes "is well ahead of the next two programming languages on this month's leaderboard: C++ is at 9.18% and C is 9.03%." But the first top six languages haven't changed since last year...

Python

C++

C

Java

C#

JavaScript

Since August of 2024 SQL has dropped from its #7 rank down to #12 (meaning Visual Basic and Go each rise up one rank from their position a year ago, into the #7 and #8 positions).

In the last year Perl has risen from the #25 position to #9, beating out Delphi/Oracle Pascal at #10, and Fortran at #11 (last year's #10). TIOBE CEO Jansen " [3]told TechRepublic in an email that many people were asking why Perl was becoming more popular, but he didn't have a definitive answer. He said he double-checked the underlying data and found the increase to be accurate, though the reason for the shift remains unclear."



[1] https://www.tiobe.com/tiobe-index/

[2] https://www.techrepublic.com/article/news-tiobe-commentary-august/

[3] https://www.techrepublic.com/article/news-tiobe-commentary-august/

ChatGPT now has nearly [1]700 million weekly users, OpenAI says. But after launching GPT-5 last week, critics bashed its less-intuitive feel, [2]reports CNBC , "ultimately leading the company to restore its legacy GPT-4 to paying chatbot customers."

Yet GPT-5 was always about cracking the enterprise market "where rival Anthropic has enjoyed a head start," they write. And one week in, "startups like Cursor, Vercel, and Factory say they've already made GPT-5 the default model in certain key products and tools, touting its faster setup, better results on complex tasks, and a lower price."

> Some companies said GPT-5 now matches or beats Claude on code and interface design, a space Anthropic once dominated. Box, another enterprise customer, has been testing GPT-5 on long, logic-heavy documents. CEO Aaron Levie told CNBC the model is a "breakthrough," saying it performs with a level of reasoning that prior systems couldn't match...

>

> Still, the economics are brutal. The models are expensive to run, and both OpenAI and Anthropic are spending big to lock in customers, with OpenAI on track to burn $8 billion this year. That's part of why both Anthropic and OpenAI are [3]courting new capital ... GPT-5 is significantly cheaper than Anthropic's top-end Claude Opus 4.1 — by a factor of seven and a half, in some cases — but OpenAI is spending huge amounts on infrastructure to sustain that edge. For OpenAI, it's a push to win customers now, get them locked in and build a real business on the back of that loyalty...

>

> GPT-5 API usage has surged since launch, with the model now processing more than twice as much coding and agent-building work, and reasoning use cases jumping more than eightfold, said a person familiar with the matter who requested anonymity in order to discuss company data. Enterprise demand is rising sharply, particularly for planning and multi-step reasoning tasks.

>

> GPT-5âs traction over the past week shows how quickly loyalties can shift when performance and price tip in OpenAI's favor. AI-powered coding platform [4]Qodo recently tested GPT-5 against top-tier models including Gemini 2.5, Claude Sonnet 4, and Grok 4, and said in a blog post that it led in catching coding mistakes. The model was often the only one to catch critical issues, such as security bugs or broken code, suggesting clean, focused fixes and skipping over code that didn't need changing, the company said. Weaknesses included occasional false positives and some redundancy.

JetBrains has also [5]adopted GPT-5 as the default for its AI Assistant and for its new no-code tool Kineto, according to the article.

But Anthropic is still enjoying a great year too, with its annualized revenue growing 17x year-over-year (according to "a person familiar with the matter who requested anonymity")



[1] https://www.cnbc.com/2025/08/04/openai-chatgpt-700-million-users.html

[2] https://www.cnbc.com/2025/08/14/gpt-5-openai-ai-enterprise.html

[3] https://www.cnbc.com/2025/07/29/anthropic-in-talks-to-raise-fresh-capital-at-170-billion-valuation.html

[4] https://www.qodo.ai/blog/benchmarking-gpt-5-on-real-world-code-reviews-with-the-pr-benchmark/

[5] https://blog.jetbrains.com/blog/2025/08/07/gpt-5-support-in-ai-assistant-junie-kineto/#

"As employers and tech companies rush to deploy AI software into workplaces to improve efficiency, labor unions are stepping up work with state lawmakers across the nation to place guardrails on its use..." [1]reports the Washington Post .

"Union leaders say they must intervene to protect workers from the potential for AI to cause massive job displacement or infringe on employment rights."

> In Massachusetts, the Teamsters labor union is backing a proposed state law that would require autonomous vehicles to have a human safety operator who can intervene during the ride, effectively forbidding truly driverless rides. Oregon lawmakers recently passed a bill supported by the Oregon Nurses Association that prohibits AI from using the title "nurse" or any associated abbreviations. The American Federation of Labor and Congress of Industrial Organizations, a federation of 63 national and international labor unions, launched a national task force last month to work with state lawmakers on more laws that regulate automation and AI affecting workers... The AFL-CIO task force plans to help unions take on problematic use of AI in collective bargaining and contracts and in coming months to develop a slate of model legislation available to state leaders, modeled on recently passed and newly proposed legislation in places including California and Massachusetts.

The president of the California Federation of Labor Unions also supports a proposed state law "that would prevent employers from primarily relying on AI software to automate decisions like terminations or disciplinary actions," according to the article. "Instead, humans would have to review decisions. The law would also prohibit use of tools that predict workers' behaviors, emotional states and personality."



[1] https://www.msn.com/en-us/news/politics/labor-unions-mobilize-to-challenge-advance-of-algorithms-in-workplaces/ar-AA1Knkkd

"Illinois last week banned the use of artificial intelligence in mental health therapy," [1]reports the Washington Post , "joining a small group of states regulating the emerging use of AI-powered chatbots for emotional support and advice."

> Licensed therapists in Illinois are now forbidden from using AI to make treatment decisions or communicate with clients, though they can still use AI for administrative tasks. Companies are also not allowed to offer AI-powered therapy services — or advertise chatbots as therapy tools — without the involvement of a licensed professional.

>

> Nevada [2]passed a similar set of restrictions on AI companies offering therapy services in June, while Utah also [3]tightened regulations for AI use in mental health in May but stopped short of banning the use of AI.

>

> The bans come as experts have raised alarms about the potential dangers of therapy with AI chatbots that haven't been reviewed by regulators for safety and effectiveness. Already, cases have emerged of chatbots engaging in [4]harmful conversations with vulnerable people — and of users revealing personal information to chatbots without realizing their conversations [5]were not private .

>

> Some AI and psychiatry experts said they welcomed legislation to limit the use of an unpredictable technology in a delicate, human-centric field.



[1] https://www.msn.com/en-us/health/other/illinois-bans-ai-therapy-as-some-states-begin-to-scrutinize-chatbots/ar-AA1KotWK

[2] https://www.leg.state.nv.us/App/NELIS/REL/83rd2025/Bill/12575/Overview

[3] https://le.utah.gov/~2025/bills/static/HB0452.html

[4] https://www.washingtonpost.com/nation/2024/10/24/character-ai-lawsuit-suicide/

[5] https://www.washingtonpost.com/technology/2025/06/13/meta-ai-privacy-users-chatbot/

Puerto Rico expects 93 different power outages this summer, [1]reports the Washington Post .

But they also note that "roughly 1 in 10 Puerto Rican homes now have a battery and solar array for backup power" which have also "become a crucial source of backup power for the entire island grid."

> A network of 69,000 home batteries can generate as much electricity as a small natural gas turbine during an emergency, temporarily covering about 2 percent of the island's energy needs when things go wrong... "It has very, very certainly prevented more widespread outages," said Daniel Haughton, [transmission and distribution planning director for Puerto Rico's grid operator]. "In the instances that we had to [cut power], it was for a much shorter duration: A four-hour outage became a one- or two-hour outage."

>

> Puerto Rico's experience offers a glimpse into the future for the rest of the United States, where batteries are starting to play a big role in keeping the lights on. Authorities in [2]Texas , [3]California and [4]New England have credited home batteries with preventing blackouts during summer energy crunches. As power grids across the country groan under the increasing strain of new data centers, factories and EVs, batteries offer a way for homeowners to protect themselves — and all of their neighbors — from the threat of outages. Batteries have been booming in the U.S. since 2022, when Congress created generous installation tax credits for homeowners and power companies.

>

> Home batteries generally come as an option alongside rooftop solar panels, according to Christopher Rauscher, head of grid services and electrification for Sunrun, a company that installs both. More than 70 percent of the people who hire Sunrun to put up solar panels also get a battery. With the tax credits — and the money saved on rising electricity costs — solar panels and batteries make financial sense for most American homes, according to [5]a study Stanford University scientists published Aug. 1 . About 60 percent of homes would save money in the long run with solar panels and batteries...

>

> Those batteries can have broader benefits, too. Utilities pay customers hundreds of dollars a year to sign their batteries up to form "virtual power plants," which send electricity to the grid whenever power plants can't keep up with demand. California's network of home batteries [6]can now add 535 megawatts of electricity in an emergency — about half as much energy as a nuclear power plant... [H]omeowners can make thousands of dollars a year lowering their energy bills, selling solar power back to the grid or enrolling their batteries in a virtual power plant, depending on their power company's policies and state regulations. "Over time, you would get the full payback for your system and basically get your backup for free," said Ram Rajagopal, an associate professor of civil and environmental engineering who co-authored the Stanford study.



[1] https://www.msn.com/en-us/news/us/home-batteries-are-saving-america-from-blackouts/ar-AA1Krtjs

[2] https://www.washingtonpost.com/climate-solutions/interactive/2024/virtual-power-plants-clean-energy-electricity-demand/

[3] https://www.canarymedia.com/articles/grid-edge/californians-saved-the-grid-again-they-should-be-paid-more-for-it?itid=lk_inline_enhanced-template

[4] https://www.canarymedia.com/articles/clean-energy/northeast-heat-wave-solar-battery-benefits

[5] https://www.nature.com/articles/s41560-025-01821-w

[6] https://www.brattle.com/wp-content/uploads/2025/08/Assessing-VPP-Performance-Impacts-of-a-Test-Event-in-California-1.pdf

Slashdot reader [1]darwinmac writes:

> Volkswagen is [2]offering a subscription model for extra horsepower on its ID.3 electric cars. Want to bump your ride from the standard 201 bhp to the full 228 bhp? That will be about £16.50 per month or £165 per year, or a one-time £649 "lifetime" fee that is tied to the car, not you. If you sell it, you have to pay again.

>

> VW [3]defended this to the BBC by saying you are basically paying for a sportier experience without buying a higher powered model upfront, calling it "nothing new." Nothing changes mechanically. You are just paying VW to essentially flip a boolean somewhere in the car's software.



[1] https://slashdot.org/~darwinmac

[2] https://www.neowin.net/news/popular-car-brand-wants-you-to-pay-monthly-to-unlock-more-horsepower/

[3] https://www.bbc.com/news/articles/c62weyp4qqgo

An anonymous reader quotes a report from the BBC:

> Drug treatment can help people newly diagnosed with ADHD (attention deficit hyperactivity disorder) to [1]reduce their risk of substance misuse, suicidal behavior, transport accidents and criminality , a study suggests. These issues are linked to common ADHD symptoms such as acting impulsively and becoming easily distracted. Some 5% of children and 2.5% of adults worldwide are thought to be affected by the disorder -- and growing numbers are being diagnosed. The findings, [2]published in the British Medical Journal (BMJ), confirm the wider potential benefits of drug treatment and could help patients decide whether to start medication, the researchers say.

The researchers found taking ADHD medication was linked to reductions of first-time instances of:

- suicidal behavior - 17%

- substance misuse - 15%

- transport accidents - 12%

- criminal behavior - 13%

When recurrent events were analyzed, the researchers found ADHD medication was linked to reductions of:

- 15% for suicide attempts

- 25% for substance misuses

- 4% for accidental injuries

- 16% for transport accidents

- 25% for criminal behavior



[1] https://www.bbc.com/news/articles/crr2j792drro

[2] https://www.bmj.com/content/390/bmj-2024-083658

[1]alternative_right shares a report from Phys.org:

> Take note of the name: ReHMGB1. A new study pinpoints this protein as being able to spread the wear and tear that comes with time as it quietly travels through the bloodstream. This adds significantly to our understanding of aging. The researchers were [2]able to identify ReHMGB1 as a critical messenger passing on the senescence signal by analyzing different types of human cells grown in the lab and conducting a variety of tests on mice. When [3]ReHMGB1 transmission was blocked in mice with muscle injuries, muscle regeneration happened more quickly, while the animals showed improved physical performance, fewer signs of cellular aging, and reduced systemic inflammation.

The findings have been [4]published in the journal Metabolism .



[1] https://slashdot.org/~alternative_right

[2] https://www.sciencealert.com/aging-can-spread-through-your-body-via-a-single-protein-study-finds

[3] https://en.wikipedia.org/wiki/HMGB1

[4] https://dx.doi.org/10.1016/j.metabol.2025.156259

Scientists in Svalbard warn Arctic glaciers are in "terminal" decline, with microbe-driven biological darkening [1]accelerating ice melt and potentially triggering major climate feedback loops . The Guardian reports:

> Recent research implicates snow and ice-dwelling microbes in positive feedback loops that can accelerate melting. With more than [2]70% of the planet's freshwater stored in ice and snow -- and billions of lives sustained by glacier-fed rivers -- this has profound implications everywhere. Yet not all polar microbes amplify global heating. Emerging evidence suggests that certain populations are -- for now -- applying a brake to methane emissions. [...] Microbes that live in surface ice and snow produce dark-colored pigments to harness sunlight and shield themselves from damaging UV light. They also trap dark-colored dust and debris. Together, these factors darken snow and ice, causing it to absorb more heat and melt faster -- a process known as "biological darkening."

>

> Microbes also respond to global changes, such as increased nutrients from air pollution, wildfire smoke or wind-blown dust from receding glaciers and expanding drylands. "The snowpack chemistry is now different to preindustrial era snow," Edwards says. Rising temperatures and longer melt seasons caused by global heating further accelerate the growth of ice-darkening microbes. Together, these factors have the potential to trigger an amplifying positive feedback loop: ice-darkening microbes nudge up temperatures and accelerate melt, exposing more nutrient-rich debris that encourage the growth of yet more microbes, which darken the surface further still.

>

> Each summer, a biologically darkened zone, visible from space, covering at least 100,000 sq km, appears on the south-western part of the Greenland ice sheet. According to a [3]2020 study , microbes there are responsible for 4.4 to 6.0-gigatons of runoff, representing up to 13% of total melt, from an ice mass that holds enough water to raise global sea levels by more than 7 meters. These effects are [4]acknowledged in IPCC reports but not yet incorporated into climate projection models. Across the European Alps, Himalayas, central Asia and beyond, at least 2 billion people depend on glacial meltwater for drinking water, agriculture and hydropower. Yet even if the world meets Paris targets, half these glaciers will not survive this century.



[1] https://www.theguardian.com/world/2025/aug/15/arctic-glaciers-face-terminal-decline-as-microbes-accelerate-ice-melt

[2] https://wmo.int/topics/cryosphere

[3] https://tc.copernicus.org/articles/14/309/2020/

[4] https://www.ipcc.ch/srocc/chapter/chapter-2/

South Korean authorities have [1]arrested the operator of Yubin Archive , a Telegram-based "pirate library" that grew to over 330,000 members by sharing textbooks, workbooks, lectures, and exam prep materials under the banner of "eliminating educational inequality." TorrentFreak reports:

> An official statement confirming the operator's arrest was published locally on August 12. The timeline suggests the arrest probably took place on or around August 9. The following notice appeared on Yubin Archive on August 11. "The Ministry of Culture and Sports' Copyright Crime Science Investigation Team used digital science investigation (forensics) and various investigation methods to identify the core operator, conduct simultaneous search and seizure at their homes, and fully secure the Telegram criminal activities," the Ministry's statement reads. "Investigations into accomplices who participated in the operation are also underway."

>

> While copyright infringement at scale is almost always a crime, regardless of content type or claimed good intention, having a Robin Hood character in the mix risks dilution of key anti-piracy messaging. No surprise then that much is being made of the existence of a 'minority room' within Yubin Archive, access to which was only permitted upon payment of a fee. "The core operator of the 'Yubin Archive', who was arrested, was found to have created a separate paid sharing channel (also known as a minority channel) while promoting the illegal sharing of learning materials as a noble act to eliminate educational inequality," the Ministry [2]notes . "In addition, the illegal sharing channel was a criminal act that could instill incorrect copyright awareness in most users, including teenagers. The Ministry of Culture and Sports is committed to continuing its efforts to track and strictly respond to illegal activities that abuse anonymous channels such as Telegram, to protect the rights of creators."



[1] https://torrentfreak.com/pirate-library-operator-arrested-study-canceled-for-330k-members-250814/

[2] https://www.mcst.go.kr/kor/s_notice/press/pressView.jsp?pSeq=21941

[1]VP.NET has released the source code for its Intel SGX enclave on [2]GitHub , allowing anyone to build the enclave and verify its mrenclave hash matches what's running on the servers. This takes "don't trust, verify" from marketing to reality, making privacy claims testable all the way down to hardware-enforced execution.

A move like this could set a new benchmark for transparency in privacy tech.



[1] https://vp.net/

[2] https://github.com/vpdotnet/vpnetd-sgx