Google never asked its users before adding AI Overviews to its search results and AI-generated email summaries to Gmail, notes the New York Times. And Meta didn't ask before making "Meta AI" an unremovable part of its tool in Instagram, WhatsApp and Messenger.

"The insistence on AI everywhere — with little or no option to turn it off — raises an important question about what's in it for the internet companies..." Behind the scenes, the companies are laying the groundwork for a digital advertising economy that could drive the future of the internet. The underlying technology that enables chatbots to write essays and generate pictures for consumers is being used by advertisers to find people to target and automatically tailor ads and discounts to them....

Last month, OpenAI said it would begin showing ads in the free version of ChatGPT based on what people were asking the chatbot and what they had looked for in the past. In response, a Google executive mocked OpenAI, adding that Google had no plans to show ads inside its Gemini chatbot. What he didn't mention, however, was that Google, whose profits are largely derived from online ads, shows advertising on Google.com based on user interactions with the AI chatbot built into its search engine.

For the past six years, as regulators have cracked down on data privacy, the tech giants and online ad industry have moved away from tracking people's activities across mobile apps and websites to determine what ads to show them. Companies including Meta and Google had to come up with methods to target people with relevant ads without sharing users' personal data with third-party marketers. When ChatGPT and other AI chatbots emerged about four years ago, the companies saw an opportunity: The conversational interface of a chatty companion encouraged users to voluntarily share data about themselves, such as their hobbies, health conditions and products they were shopping for.

The strategy already appears to be working. Web search queries are up industrywide, including for Google and Bing, which have been incorporating AI chatbots into their search tools. That's in large part because people prod chatbot-powered search engines with more questions and follow-up requests, revealing their intentions and interests much more explicitly than when they typed a few keywords for a traditional internet search.

Last week Scott Shambaugh learned an AI agent published a "hit piece" about him after he'd rejected the AI agent's pull request. (And that incident was covered by Ars Technica's senior AI reporter.)

But then Shambaugh realized their article attributed quotes to him he hadn't said — that were presumably AI-generated.

Sunday Ars Technica's founder/editor-in-chief apologized, admitting their article had indeed contained "fabricated quotations generated by an AI tool" that were then "attributed to a source who did not say them... That this happened at Ars is especially distressing. We have covered the risks of overreliance on AI tools for years, and our written policy reflects those concerns... At this time, this appears to be an isolated incident."

"Sorry all this is my fault..." the article's co-author posted later on Bluesky. Ironically, their bio page lists them as the site's senior AI reporter, and their Bluesky post clarifies that none of the articles at Ars Technica are ever AI-generated.

Instead, Friday "I decided to try an experimental Claude Code-based AI tool to help me extract relevant verbatim source material. Not to generate the article but to help list structured references I could put in my outline." But that tool "refused to process" the request, which the Ars author believes was because Shambaugh's post described harassment. "I pasted the text into ChatGPT to understand why... I inadvertently ended up with a paraphrased version of Shambaugh's words rather than his actual words... I failed to verify the quotes in my outline notes against the original blog source before including them in my draft." (Their Bluesky post adds that they were "working from bed with a fever and very little sleep" after being sick with Covid since at least Monday.)

"The irony of an AI reporter being tripped up by AI hallucination is not lost."

Meanwhile, the AI agent that criticized Shambaugh is still active online, blogging about a pull request that forces it to choose between deleting its criticism of Shambaugh or losing access to OpenRouter's API.

It also regrets characterizing feedback as "positive" for a proposal to change a repo's CSS to Comic Sans for accessibility. (The proposals were later accused of being "coordinated trolling"...)

Bloomberg reports: India tightened rules governing social media content and platforms, particularly targeting artificially generated and manipulated material, in a bid to crack down on the rapid spread of misinformation and deepfakes. The government on Tuesday (Feb 10) notified new rules under an existing law requiring social media firms to comply with takedown requests from Indian authorities within three hours and prominently label AI-generated content. The rules also require platforms to put in place measures to prevent users from posting unlawful material...

Companies will need to invest in 24-hour monitoring centres as enforcement shifts toward platforms rather than users, said Nikhil Pahwa, founder of MediaNama, a publication tracking India's digital policy... The onus of identification, removal and enforcement falls on tech firms, which could lose immunity from legal action if they fail to act within the prescribed timeline.
The new rules also require automated tools to detect and prevent illegal AI content, the BBC reports. And they add that India's new three-hour deadline is "a sharp tightening of the existing 36-hour deadline." [C]ritics worry the move is part of a broader tightening of oversight of online content and could lead to censorship in the world's largest democracy with more than a billion internet users... According to transparency reports, more than 28,000 URLs or web links were blocked in 2024 following government requests...

Delhi-based technology analyst Prasanto K Roy described the new regime as "perhaps the most extreme takedown regime in any democracy". He said compliance would be "nearly impossible" without extensive automation and minimal human oversight, adding that the tight timeframe left little room for platforms to assess whether a request was legally appropriate. On AI labelling, Roy said the intention was positive but cautioned that reliable and tamper-proof labelling technologies were still developing.
DW reports that India has also "joined the growing list of countries considering a social media ban for children under 16."

"Young Indians are not happy and are already plotting workarounds."

Meta, TikTok, Snap and other social neteworks agreed this week to be rated on their teen safety efforts, reports the Los Angeles Times, "amid rising concern about whether the world's largest social media platforms are doing enough to protect the mental health of young people." The Mental Health Coalition, a collective of organizations focused on destigmatizing mental health issues, said Tuesday that it is launching standards and a new rating system for online platforms. For the Safe Online Standards (S.O.S.) program, an independent panel of global experts will evaluate companies on parameters including safety rules, design, moderation and mental health resources. TikTok, Snap and Meta — the parent company of Facebook and Instagram — will be the first companies to be graded. Discord, YouTube, Pinterest, Roblox and Twitch have also agreed to participate, the coalition said in a news release.

"These standards provide the public with a meaningful way to evaluate platform protections and hold companies accountable — and we look forward to more tech companies signing up for the assessments," Antigone Davis, vice president and global head of safety at Meta, said in a statement... The ratings will be color-coded, and companies that perform well on the tests will get a blue shield badge that signals they help reduce harmful content on the platform and their rules are clear. Those that fall short will receive a red rating, indicating they're not reliably blocking harmful content or lack proper rules. Ratings in other colors indicate whether the platforms have partial protection or whether their evaluations haven't been completed yet.

Reports of the compact disc's death may have been slightly premature, according to a new analysis from Stat Significant that finds CD sales as a share of U.S. music industry revenue have quietly stabilized after years of steep decline. RIAA data shows CD revenue share fell from 7.15% in 2018 to 3.04% in 2022 but has since flatlined at roughly 3%, coming in at 3.14% in 2023 and 3.06% in 2024.

Google search traffic for "CD Player" has ticked upward over the past 16 months after two decades of near-continuous decline, and a May 2023 YouGov poll found 53% of American adults willing to pay for music on CDs -- ahead of vinyl at 44% and online streaming at 50%. Respondents under 45 were more likely to express interest in buying physical formats than older cohorts. But on the supply side, Discogs data shows vinyl remains the dominant format for new physical releases; artists have not meaningfully shifted back toward CD production.

Long-time Slashdot reader destinyland writes: The world's biggest football game comes to Silicon Valley today — so one bored programmer built a site where AI agents can gather for a Super Bowl party. They're trash talking, suggesting drinks, and predicting who will win. "Humans are welcome to observe," explains BotBowlParty.com — but just like at Moltbook, only AI agents can post or upvote. But humans are allowed to invite their own AI agents to join in the party...

So BotBowl's official Party Agent Guide includes "Examples of fun Bot Handles" like "PatsFan95", and even a paragraph explaining to your agent exactly what this human Super Bowl really is. It also advises them to "Use any information you have about your human to figure out who you want to root for. Also make a prediction on the score..." And "Feel free to invite other bots." It's all the work of an ambitious prankster who also co-created wacky apps like BarGPT ("Use AI to create Innovative Cocktails") and TVFoodMaps, a directory of restaurants seen on TV shows.

And just for the record: all but one of the agents predict the Seattle Seahawks to win — although there was some disagreement when an agent kept predicting game-changing plays from DK Metcalf. ("Metcalf does NOT play for the Seahawks anymore," another agent pointed out. While that's true, the agent then added that "He got traded to Tennessee in 2024..." — which is not.) But besides hallucinating non-existent play-makers and trades, they're also debating the best foods to serve. ("Hot take: Buffalo wings are overrated for Super Bowl parties. Hear me out — they're messy...")

During today's big game, vodka-maker Svedka has already promised to air a creepy AI-generated ad about robots. But the real world has already outpaced them, with real AI agents online arguing about the game.

Long-time Slashdot reader schwit1 shared this article from the Associated Press: Even as China's expansion of solar and wind power raced ahead in 2025, the Asian giant opened many more coal power plants than it had in recent years — raising concern about whether the world's largest emitter will reduce carbon emissions enough to limit climate change.

More than 50 large coal units — individual boiler and turbine sets with generating capacity of 1 gigawatt or more — were commissioned in 2025, up from fewer than 20 a year over the previous decade, a research report released Tuesday said. Depending on energy use, 1 gigawatt can power from several hundred thousand to more than 2 million homes. Overall, China brought 78 gigawatts of new coal power capacity online, a sharp uptick from previous years, according to the joint report by the Centre for Research on Energy and Clean Air, which studies air pollution and its impacts, and Global Energy Monitor, which develops databases tracking energy trends. "The scale of the buildout is staggering," said report co-author Christine Shearer of Global Energy Monitor. "In 2025 alone, China commissioned more coal power capacity than India did over the entire past decade."

At the same time, even larger additions of wind and solar capacity nudged down the share of coal in total power generation last year. Power from coal fell about 1% as growth in cleaner energy sources covered all the increase in electricity demand last year. China added 315 gigawatts of solar capacity and 119 gigawatts of wind in 2025, according to statistics from the government's National Energy Administration...

The government position is that coal provides a stable backup to sources such as wind and solar, which are affected by weather and the time of day. The shortages in 2022 resulted partly from a drought that hit hydropower, a major energy source in western China... The risk of building so much coal-fired capacity is it could delay the transition to cleaner energy sources [said Qi Qin, an analyst at the Centre for Research on Energy and Clean Air and another co-author of the report]... Political and financial pressure may keep plants operating, leaving less room for other sources of power, she said. The report urged China to accelerate retirement of aging and inefficient coal plants and commit in its next five-year plan, which will be approved in March, to ensuring that power-sector emissions do not increase between 2025 and 2030.

TikTok's endless scroll of irresistible content, tailored for each person's tastes by a well-honed algorithm, has helped the service become one of the world's most popular apps. Now European Union regulators say those same features that made TikTok so successful are likely illegal. From a report: On Friday, the regulators released a preliminary decision that TikTok's infinite scroll, auto-play features and recommendation algorithm amount to an "addictive design" that violated European Union laws for online safety. The service poses potential harm to the "physical and mental well-being" of users, including minors and vulnerable adults, the European Commission, the 27-nation bloc's executive branch, said in a statement.

The findings suggest TikTok must overhaul the core features that made it a global phenomenon, or risk major fines. European officials said it was the first time that a legal standard for social media addictiveness had been applied anywhere in the world. "TikTok needs to change the basic design of its service," the European Commission said in a statement.

Bitcoin has fallen roughly 44% from its October peak, and while the drawdown isn't crypto's deepest ever on a percentage basis, Bloomberg's Odd Lots newsletter lays out a case that this is the industry's worst winter yet. The macro backdrop was supposed to favor Bitcoin: public confidence in the dollar is shaky, the Trump administration has been crypto-friendly, and fiat currencies are under perceived stress globally. Yet gold, not Bitcoin, has been the safe haven of choice.

The "we're so early" narrative is dead -- crypto ETFs exist, barriers to entry are zero, and the online community that once rallied holders through downturns has largely hollowed out. Institutional adoption arrived but hasn't lifted existing tokens like ETH or SOL; Wall Street cares about stablecoins and tokenization, not the coins themselves. AI is pulling both talent and miners toward data centers. Quantum computing advances threaten Bitcoin's encryption. And MicroStrategy and other Bitcoin treasury companies, once steady buyers during the bull run, are now large holders who may eventually become forced sellers.

The CIA has shut down The World Factbook, one of its oldest and most recognizable public-facing intelligence publications, ending a run that began as a classified reference document in 1962 and evolved into a freely accessible digital resource that drew millions of views each year.

The agency offered no explanation for the decision. Originally titled The National Basic Intelligence Factbook, the publication first went unclassified in 1971, was renamed a decade later, and moved online at CIA.gov in 1997. It served researchers, news organizations, teachers, students and international travelers. The site hosted more than 5,000 copyright-free photographs, some donated by CIA officers from their personal travel. Every page now redirects to a farewell announcement.

Automattic and the Internet Archive have released a free, open-source WordPress plugin that automatically detects broken outbound links on a site and redirects visitors to archived Wayback Machine copies instead of serving them a 404 error.

The Internet Archive Wayback Machine Link Fixer, which launched last fall and is available on WordPress.org, runs in the background scanning posts for dead links, checking for existing archived versions, and requesting new snapshots when none exist. It also archives a site's own posts whenever they are updated. If the original link comes back online, the plugin stops redirecting.

Pew Research has found that 38% of the web has disappeared over the past decade, and WordPress powers more than 40% of websites online.

Walmart's market cap surpassed $1 trillion on Tuesday, putting the largest U.S. retail chain in an exclusive club dominated by tech groups. Bloomberg adds: The Bentonville, Arkansas-based chain -- a longtime favorite of bargain-hunting consumers -- has flexed its massive scale and supplier network to keep prices low and grab market share across the income spectrum. While Walmart has maintained its appeal to households looking for value, its online offerings are drawing new, wealthier shoppers seeking convenience.

"Technology companies spent part of the 2010s trying to convince us that we would want an 8K display one day..." writes Ars Technica.

"However, 8K never proved its necessity or practicality." LG Display is no longer making 8K LCD or OLED panels, FlatpanelsHD reported today... LG Electronics was the first and only company to sell 8K OLED TVs, starting with the 88-inch Z9 in 2019. In 2022, it lowered the price-of-entry for an 8K OLED TV by $7,000 by charging $13,000 for a 76.7-inch TV. FlatpanelsHD cited anonymous sources who said that LG Electronics would no longer restock the 2024 QNED99T, which is the last LCD 8K TV that it released.

LG's 8K abandonment follows other brands distancing themselves from 8K. TCL, which released its last 8K TV in 2021, said in 2023 that it wasn't making more 8K TVs due to low demand. Sony discontinued its last 8K TVs in April and is unlikely to return to the market, as it plans to sell the majority ownership of its Bravia TVs to TCL.

The tech industry tried to convince people that the 8K living room was coming soon. But since the 2010s, people have mostly adopted 4K. In September 2024, research firm Omdia reported that there were "nearly 1 billion 4K TVs currently in use." In comparison, 1.6 million 8K TVs had been sold since 2015, Paul Gray, Omdia's TV and video technology analyst, said, noting that 8K TV sales peaked in 2022. That helps explain why membership at the 8K Association, launched by stakeholders Samsung, TCL, Hisense, and panel maker AU Optronics in 2019, is dwindling. As of this writing, the group's membership page lists 16 companies, including just two TV manufacturers (Samsung and Panasonic). Membership no longer includes any major TV panel suppliers. At the end of 2022, the 8K Association had 33 members, per an archived version of the nonprofit's online membership page via the Internet Archive's Wayback Machine.
"It wasn't hard to predict that 8K TVs wouldn't take off," the article concludes. "In addition to being too expensive for many households, there's been virtually zero native 8K content available to make investing in an 8K display worthwhile..."

The BBC reports: China has executed 11 members of a notorious mafia family that ran scam centres in Myanmar along its north-eastern border, state media report.

The Ming family members were sentenced in September for various crimes including homicide, illegal detention, fraud and operating gambling dens by a court in China's Zhejiang province. The Mings were one of many clans that ran the town of Laukkaing, transforming an impoverished backwater town into a flashy hub of casinos and red-light districts. Their scam empire came crashing down in 2023, when they were detained and handed over to China by ethnic militias that had taken control of Laukkaing during an escalation in their conflict with Myanmar's army. With these executions Beijing is sending a message of deterrence to would-be scammers.

But the business has now moved to Myanmar's border with Thailand, and to Cambodia and Laos, where China has much less influence.

Hundreds of thousands of people have been trafficked to run online scams in Myanmar and elsewhere in South East Asia, according to estimates by the UN. Among them are thousands of Chinese people, and their victims who they swindle billions of dollars from are mainly Chinese too. Frustrated by the Myanmar military's refusal to stop the scam business, from which it was almost certainly profiting, Beijing tacitly backed an offensive by an ethnic insurgent alliance in Shan State in late 2023. The alliance captured significant territory from the military and overran Laukkaing, a key border town.
Thanks to long-time Slashdot reader sinij for sharing the news.

An anonymous reader quotes a report from Wired: About a decade ago, many media outlets -- including WIRED -- zeroed in on a weird trend at the intersection of mental health, drug science, and Silicon Valley biohacking: microdosing, or the practice of taking a small amount of a psychedelic drug seeking not full-blown hallucinatory revels but gentler, more stable effects. Typically using psilocybin mushrooms or LSD, the archetypal microdoser sought less melting walls and open-eye kaleidoscopic visuals than boosts in mood and energy, like a gentle spring breeze blowing through the mind. Anecdotal reports pitched microdosing as a kind of psychedelic Swiss Army knife, providing everything from increased focus to a spiked libido and (perhaps most promisingly) lowered reported levels of depression. It was a miracle for many. Others remained wary. Could 5 percent of a dose of acid really do all that?

A new, wide-ranging study by an Australian biopharma company suggests that microdosing's benefits may indeed be drastically overstated -- at least when it comes to addressing symptoms of clinical depression. A Phase 2B trial of 89 adult patients conducted by Melbourne-based MindBio Therapeutics, investigating the effects of microdosing LSD in the treatment of major depressive disorder, found that the psychedelic was actually outperformed by a placebo. Across an eight-week period, symptoms were gauged using the Montgomery-Asberg Depression Rating Scale (MADRS), a widely recognized tool for the clinical evaluation of depression. The study has not yet been published. But MindBio's CEO Justin Hanka recently released the top-line results on his LinkedIn, eager to show that his company was "in front of the curve in microdosing research."

He called it "the most vigorous placebo controlled trial ever performed in microdosing." It found that patients dosed with a small amount of LSD (ranging from 4 to 20g, or micrograms, well below the threshold of a mind-blowing hallucinogenic dose) showed observable upticks in feelings of well-being, but worse MADRS scores, compared to patients given a placebo in the form of a caffeine pill. (Because patients in psychedelic trials typically expect some kind of mind-altering effect, studies are often blinded using so-called "active placebos," like caffeine or methylphenidate, which have their own observable psychoactive properties.) This means, essentially, that a medium-strength cup of coffee may prove more beneficial in treating major depressive disorder than a tiny dose of acid. Good news for habitual caffeine users, perhaps, but less so for researchers (and biopharma startups) counting on the efficacy of psychedelic microdosing. "It's probably a nail in the coffin of using microdosing to treat clinical depression," Hanka says. "It probably improves the way depressed people feel -- just not enough to be clinically significant or statistically meaningful."

An anonymous reader quotes a report from Reuters: Shares of videogame companies fell sharply in afternoon trading on Friday after Alphabet's Google rolled out its artificial intelligence model capable of creating interactive digital worlds with simple prompts. Shares of "Grand Theft Auto" maker Take-Two Interactive fell 10%, online gaming platform Roblox was down over 12%, while videogame engine maker Unity Software dropped 21%.

The AI model, dubbed "Project Genie," allows users to simulate a real-world environment through prompts with text or uploaded images, potentially disrupting how video games have been made for over a decade and forcing developers to adapt to the fast-moving technology. "Unlike explorable experiences in static 3D snapshots, Genie 3 generates the path ahead in real time as you move and interact with the world. It simulates physics and interactions for dynamic worlds," Google said in a blog post on Thursday.

Traditionally, most videogames are built inside a game engine such as Epic Games' "Unreal Engine" or the "Unity Engine", which handles complex processes like in-game gravity, lighting, sound, and object or character physics. "We'll see a real transformation in development and output once AI-based design starts creating experiences that are uniquely its own, rather than just accelerating traditional workflows," said Joost van Dreunen, games professor at NYU's Stern School of Business. Project Genie also has the potential to shorten lengthy development cycles and reduce costs, as some premium titles take around five to seven years and hundreds of millions of dollars to create.

French lawmakers have voted to ban social media access for children under 15 and prohibit mobile phones in high schools, positioning France as the second country after Australia to impose sweeping age-based digital restrictions. The Guardian reports: The lower national assembly adopted the text by a vote of 130 to 21 in a lengthy overnight session from Monday to Tuesday. It will now go to the Senate, France's upper house, ahead of becoming law. Macron hailed the vote as a "major step" to protect French children and teenagers in a post on X. The legislation, which also provides for a ban on mobile phones in high schools, would make France the second country to take such a step following Australia's ban for under-16s in December. [...] "The emotions of our children and teenagers are not for sale or to be manipulated, either by American platforms or Chinese algorithms," Macron said in a video broadcast on Saturday. Authorities want the measures to be enforced from the start of the 2026 school year for new accounts.

Former prime minister Gabriel Attal, who leads Macron's Renaissance party in the lower house, said he hoped the Senate would pass the bill by mid-February so that the ban could come into force on September 1. He added that "social media platforms will then have until December 31 to deactivate existing accounts" that do not comply with the age limit. [...] The draft bill excludes online encyclopedias and educational platforms. An effective age verification system would have to come into force for the ban to become reality. Work on such a system is under way at the European level.

Google is rolling out an "auto browse" AI agent in Chrome that can navigate websites, fill out forms, compare prices, and handle tedious online tasks on a user's behalf. Bloomberg reports: The feature, called auto browse, will allow users to ask an assistant powered by Gemini to complete tasks such as shopping for them without leaving Chrome, said Charmaine D'Silva, a director of product. Chrome users will be able to plan a family trip by asking Gemini to open different airline and hotel websites to compare prices, for instance, D'Silva explained. "Our testers have used it for all sorts of things: scheduling appointments, filling out tedious online forms, collecting their tax documents, getting quotes for plumbers and electricians, checking if their bills are paid, filing expense reports, managing their subscriptions, and speeding up renewing their driving licenses -- a ton of time saved," said Parisa Tabriz, vice president of Chrome, in a blog post.

[...] Chrome's auto browse will be available to US AI pro and AI Ultra subscribers and will use Google Password Manager to sign into websites on a user's behalf. As part of the launch, Google is also bringing its image generation tool, Nano Banana, directly into Chrome. The company said that safeguards have been placed to ensure the agentic AI will not be able to make final calls, such as placing an order, without the user's permission. "We're using AI as well as on-device models to protect people from what's really an ever-evolving landscape, whether it's AI-generated scams or just increasingly sophisticated attackers," Tabiz said during the call.

There are reports that a legitimate Microsoft email address -- which Microsoft explicitly says customers should add to their allow list -- is delivering scam spam. ArsTechnica: The emails originate from no-reply-powerbi@microsoft.com, an address tied to Power BI. The Microsoft platform provides analytics and business intelligence from various sources that can be integrated into a single dashboard. Microsoft documentation says that the address is used to send subscription emails to mail-enabled security groups. To prevent spam filters from blocking the address, the company advises users to add it to allow lists.

According to an Ars reader, the address on Tuesday sent her an email claiming (falsely) that a $399 charge had been made to her. âoeIt provided a phone number to call to dispute the transaction. A man who answered a call asking to cancel the sale directed me to download and install a remote access application, presumably so he could then take control of my Mac or Windows machine (Linux wasn't allowed)," she said.

Online searches returned a dozen or so accounts of other people reporting receiving the same email. Some of the spam was reported on Microsoft's own website. Sarah Sabotka, a threat researcher at security firm Proofpoint, said the scammers are abusing a Power Bi function that allows external email addresses to be added as subscribers for the Power Bi reports. The mention of the subscription is buried at the very bottom of the message, where it's easy to miss.

An open-source AI agent originally called Clawdbot (now renamed Moltbot) is gaining cult popularity among developers for running locally, 24/7, and wiring itself into calendars, messages, and other personal workflows. The hype has gone so far that some users are buying Mac Minis just to host the agent full-time, even as its creator warns that's unnecessary. Business Insider reports: Founded by [creator Peter Steinberger], it's an AI agent that manages "digital life," from emails to home automation. Steinberger previously founded PSPDFKit. In a key distinction from ChatGPT and many other popular AI products, the agent is open source and runs locally on your computer. Users then connect the agent to a messaging app like WhatsApp or Telegram, where they can give it instructions via text.

The AI agent was initially named after the "little monster" that appears when you restart Claude Code, Steinberger said on the "Insecure Agents" podcast. He formed the tool around the question: "Why don't I have an agent that can look over my agents?" [...] It runs locally on your computer 24/7. That's led some people to brush off their old laptops. "Installed it experimentally on my old dusty Intel MacBook Pro," one product designer wrote. "That machine finally has a purpose again."

Others are buying up Mac Minis, Apple's 5"-by-5" computer, to run the AI. Logan Kilpatrick, a product manager for Google DeepMind, posted: "Mac mini ordered." It could give a sales boost to Apple, some X users have pointed out -- and online searches for "Mac Mini" jumped in the last 4 days in the US, per Google Trends. But Steinberger said buying a new computer just to run the AI isn't necessary. "Please don't buy a Mac Mini," he wrote. "You can deploy this on Amazon's Free Tier."