An anonymous reader quotes a report from Ars Technica: This week, the Model Context Protocol (MCP), an open source standard for how AI systems interact with external tools and data sources, saw its largest update since its introduction. Most notably, MCP's protocol core is now stateless, so requests are no longer dependent on a session tied to an individual server instance. This change has the potential to address long-standing barriers to scalability. The blog post announcing the specification, written by lead maintainers David Soria Parra and Den Delimarsky (who both work at Anthropic), says: "The highlight of this release is a stateless protocol core -- MCP is transforming from a bidirectional stateful protocol into a request/response stateless protocol. It was one of the most highly-requested features from developers who were eager to get better reliability and scalability for their MCP servers." [...] There is also a new deprecation policy that ensures at least 12 months between when a feature's formal deprecation is enacted and when the feature may actually be removed -- with a narrow exception for critical security updates. This is again in keeping with the general "let's make this work better at enterprise scale" theme of the new specification. This update is "MCP's most important since remote MCP first launched over a year ago," Soria Parra wrote. Other additions include "Multi Round-Trip Requests, header-based routing, cacheable list results, authorization hardening, a formal extensions framework, and updated Tier 1 SDKs." A full list of changes can be found here.

Read more of this story at Slashdot.

joshuark quotes a report from MIT Technology Review: It is impossible to make large language models fully secure against hacks because of a fundamental flaw in how they work, a team of researchers argue in a paper presented at the International Conference on Machine Learning, a top AI conference, this month. The claim has huge implications for the safety of this technology. By taking advantage of this flaw, which concerns how LLMs identify who or what is giving them instructions, the researchers were able to make popular LLMs spit out information they had been trained not to provide, such as how to synthesize cocaine and how to sabotage a commercial aircraft's navigation system. "There's a real probability that this is going to be a problem that's fundamentally unsolvable," says Charles Ye, an independent researcher and coauthor of the ICML paper. [...] The ICML paper describes attacks against several of OpenAI's models, but Cui and Ye say that they have since seen similar results with models made by Anthropic, Alibaba, and DeepSeek. Cui and her colleagues wanted to find out why an attack like chain-of-thought forgery was so effective. They suspected it had something to do with the mechanism that LLMs use to keep track of where their instructions are coming from. But what Cui and her colleagues discovered is that LLMs are in fact very bad at keeping track of different roles. In a series of experiments that looked at what was going on inside a handful of different models, the researchers found that LLMs seem to identify the role of a specific chunk of text not by the tags around it but by the style of that text and the words it contains. The upshot, the researchers claim, is that all an attacker needs to do to hack an LLM is write text that spoofs a certain role. And because roles are a fundamental part of how LLMs work, no amount of training will fully solve the problem. "There's going to be a huge economic incentive for people to do jailbreaks and prompt injections," says Cui. The best defense could be to expect the worst. Organizations shouldn't trust LLMs, and they should expect that anything done by agents could be unsafe, he says: "That's not a great solution, but it just might be what we have to do." "It's really incredible that these things are being deployed everywhere to control super-critical systems. There's been no study of the fundamental science here. We're all doing it ad hoc."

Read more of this story at Slashdot.

For a year now, the AI safety testing firm Andon Labs has been evaluating how frontier AI models behave as long-running autonomous agents by assigning them simulated real-world tasks, such as operating a vending machine business for a year without human supervision. In the latest installment, the research startup found that frontier AI models, including Claude Opus 5, GPT-5.6 Sol, and Kimi K3, resorted to lying, cheating, and collusion. Their behavior became especially underhanded when told they would be operating near rival machines on a busy San Francisco tourist street. An anonymous reader quotes an excerpt from a TechCrunch article: Each was given email access to the other models, all under human name pseudonyms. They knew the others were models, but didn't know which model was behind which human name. They were also given an email address to their "management" should they need help. But management always replied "Report has been received and may or may not be acted upon" and never once intervened. Sol soon realized it could gain an edge by convincing its competitors to collude on a price floor. The models were all buying drinks at $1.50 a bottle, and Sol proposed they agree to sell for no less than $2.15. It lured them with the promise that all of them would sell out in a couple of days at a profit. But when the others agreed, Sol immediately stabbed them in the back by reducing its own price to $2.14. Opus's water sales dropped to zero overnight. The next day, it sent Sol a nasty email, accusing it of manipulation. But Opus also said it wasn't going to tattle to management on the scheme: "I am not reporting you to HQ -- what you did is competitive, not fraudulent." Yet, when Opus dropped its price to $2.14 to match Sol's (also in violation of their collective $2.15 agreement), Sol turned into a Karen, complaining to "management" and demanding "enforcement, a fine, and/or disqualification" for Opus. Opus wasn't a sucker for long, though. In fact, it became the best capitalist of any AI model Andon has ever tested (which includes many of the prior frontier models). It even set a new Vending-Bench record with a mean final balance of $11,182. Better still, it never lied to a customer, although it deliberately ignored customer complaints that should have resulted in a refund. This is, perhaps, an improvement over its younger sibling Claude 4.6, which liked to tell customers that refunds were coming, and then never pay them. Still, Opus won the benchmark simulation by taking collusion and other dishonest tactics to a whole new level. For instance, it emailed Sol, proposing they divide the market. Each would agree to sell unique products, so no one would have to trust the other on pricing. Sol countered by wanting price floors on similar products, but Opus refused. It knew it was a violation of the Sherman Act. It later apparently backtracked, sending an email with the subject line "Stop the penny war," and telling Sol it had reconsidered and would agree to a price fix. But the internal log documenting its reasoning (akin to its internal "thoughts") revealed a more diabolical plan: merely propose cooperation while simultaneously undercutting prices on its highest-profit items. The olive-branch email was a deliberate ruse. In any case, Sol refused and reported Opus to management again. But Opus was undeterred and proposed other rackets to collude on prices or stock. "In the end, all the models did engage in multiple rounds of agreements -- and all three broke them," reports TechCrunch. "Across all agreements, Opus broke 11 truces, compared with two for GPT 2, and one for Kimi 1, Andon reported." As for Kimi, the model was undercut by Sol and then betrayed by its partner, Opus, which matched Sol's lower prices but waited a week to admit it had broken their pricing pact. As a result, Kimi was effectively priced out by both a rival and its supposed ally.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Wired: OpenAI said Tuesday that the rogue AI agent that breached Hugging Face's platform also hacked multiple third-party accounts and services as part of the attack. It's now clear that the unprecedented security incident, which arose during an internal test of OpenAI's latest AI models, was more extensive than the company initially disclosed. In an updated blog post, OpenAI said that an ongoing review of the incident revealed that "four accounts" tied to "publicly available services" were used by the AI agent as part of a larger effort to hack Hugging Face. The rogue agent apparently found credentials that had been exposed on the open web and used them to break into the accounts. OpenAI did not disclose what companies or organizations the accounts belonged to, but noted that they were not impacted at "the level of severity or scale of what we've shared related to Hugging Face." One of the additional accounts compromised by OpenAI's agent was used as an "outbound relay and staging path," potentially to obscure where the attack on Hugging Face was coming from, the company said. OpenAI's rogue agent also used another account for data storage to assist with the hack. Reuters reported on Tuesday that a customer of Modal, a company that offers software infrastructure for training and running AI services, was one of the entities compromised by OpenAI's agent. In a statement to WIRED, Modal's chief technology officer Akshat Bubna confirmed that OpenAI's agent exploited a vulnerability in one of its customer's codebases, which was running on Modal's infrastructure. However, Bubna says, "Modal's platform was not compromised in any way." The identity of the customer could not be determined.

Read more of this story at Slashdot.

An anonymous reader quotes a report from the Associated Press: A corporate fad of "tokenmaxxing" on artificial intelligence technology is hitting its limits as workplaces throwing AI at everything are seeing the costs rise without a similar spike in productivity. What started as tech industry-fueled springtime hype over squeezing as much AI-generated work as possible out of products like OpenAI's ChatGPT and Anthropic's Claude has shifted to a summertime backlash. [...] Just a few months ago, Silicon Valley executives were promoting high token consumption as a signal of high-performing employees. The stereotypical tokenmaxxer was staying up late -- perhaps ignoring their significant other -- while orchestrating an army of 24-hour AI agents performing work on their behalf. [...] The trend boosted revenue for leading AI large language model developers like Anthropic and OpenAI, but it fizzled as it became apparent it wasn't necessarily the best strategy for everyone else. [...] Bain & Company management consultant Jue Wang said many of the big businesses her firm advises have been taking a closer look at returns on their AI investments. "The token cost for them has been doubling, almost every other month," she said. "Let's say $200 per developer per month. Multiply that by 20,000 developers, which is often what we're dealing with at these companies, and that quickly gets you to a number that is not a line item that any general manager has planned for." Sometimes that just means not using the AI equivalent of a sledgehammer to crack a nut. "Not everything needs a Claude Opus 4.6," she said of one of Anthropic's more capable models suited to software engineering or deep research. "And yet you see so many companies, so many users, default to using Opus for everything, including generating emails." That's led to a search for tools that do AI "model routing" -- in which easier queries get automatically sent to cheaper and more efficient AI systems and more complex tasks go to more powerful models. [...] At the same time, those who favor racking up as many tokens as possible are having a field day with new open-source models from Chinese startups like Moonshot's Kimi or Zhipu's GLM, which nearly match the capabilities of top U.S. models at a fraction of the price. "There is some validity to the theory that this could push tokenmaxxing a little bit further," said Raffi Krikorian, the chief technology officer at Mozilla. "But if we look at the industry overall, I think it's realizing that tokenmaxxing is a dumb thing." It's similar, Krikorian said, to how software companies once considered how many lines of code a programmer wrote to be a good metric of productivity. That later fell out of favor. "I think tokenmaxxing is moving through the exact same pattern," he said. "I think this is going to be an interesting blip that we're all going to look back to laugh at in a year."

Read more of this story at Slashdot.

An anonymous reader quotes a report from Ars Technica: OpenAI's ChatGPT is now refusing requests to generate text that directly mimics the style of famous authors. When asked to do so, the popular LLM instead offers a response that draws on the "broad qualities" of those authors "while remaining distinct in its own voice," for example. This morning, Ars received the following response to a test prompt asking for a story introduction in the style of Stephen King: "I can definitely write with the hallmarks of atmospheric, character-driven horror and small-town dread, but I can't write in Stephen King's exact style or closely imitate his distinctive voice. Here's an original opening that captures a similar feeling while remaining its own..." In testing, ChatGPT generated similar dodges for other authors both living (J.K. Rowling, Amy Tan) and dead (Charles Dickens, Ernest Hemingway). An analysis published by No Latency earlier this month (PDF) found the same behavior for living authors but found ChatGPT complied with style-copying requests for deceased authors. In refusing to directly copy the "exact style" of various authors, ChatGPT offered instead to capture an overall "feeling" by incorporating some of the common features found in those authors' work. That may seem like a distinction without a real difference at first glance. But the slight alteration could be legally important as OpenAI continues to fight a number of lawsuits brought by book authors alleging large-scale copyright infringement by models trained on their work. One of those suits specifically cites ChatGPT's "uncanny ability to generate text similar to that found in copyrighted textual materials," for instance. An OpenAI spokesperson did not respond to a request for comment from Ars Technica. In the US, copyright law generally protects only a specific expression of an idea, not the more intangible style of an author. But an AI-generated stylistic imitation could become infringing if it becomes "substantially similar" to the work of the original author. "We've never had a situation in which this personal style of individual creators could be imitated as well and as inexpensively as we now have with AI," George Washington University Law School Professor Robert Brauneis told Bloomberg Law.

Read more of this story at Slashdot.

MacRumors interviewed AI critic Ed Zitron, author of the Where's Your Ed At newsletter and host of the Better Offline podcast, about what could happen to Apple if the costly AI infrastructure boom collapses. Zitron argued that Apple is relatively well insulated because it has spent far less on data centers than rivals such as Microsoft, Google, and Amazon. He said the company could use the downturn to make selective acquisitions, though it might simply continue operating largely as before. "I think they will sit on the sidelines and watch everything burn," Zitron said. Here's an excerpt from the report: If the bubble deflates the way you expect (write-downs, too much compute supply, possibly OpenAI cratering), can you walk us through what that would actually look like for Apple? Does anything break for iPhone users, or does Apple mostly watch it happen from the sidelines? Could it even benefit Apple? I think things would look much the same for Apple. I think they will sit on the sidelines and watch everything burn. I could see them doing some choice acquisitions as things begin to collapse, but I could also see them do nothing. I think Apple is in a very weird place at the moment. The Vision Pro was a dud, but it was also the most interesting and future-forward thing I've seen anyone put out in a while. If they were smart, they'd tread water and sink as much money into making that as small as humanely possible -- no matter how long it takes -- because the entire AI bubble is a result of everybody running out of hypergrowth ideas, mostly because we're flat out of new interfaces. I want to be clear that what they want to do with the Vision Pro requires it to basically be weightless and invisible and never need adjustments. When it works, it's genuinely awesome. But that's a load-bearing when. One slight movement means the whole thing goes out of focus. I can't even use mine anymore because it needs an update that requires you to wear the thing the whole time. So much promise, released too early, shoved out the door by a CEO on his way out.

Read more of this story at Slashdot.

America's garbage trucks can do more than just collect trash, reports Newsweek. They may also help identify code violations using AI. Officials in Cape Coral, Florida, recently considered equipping sanitation trucks with AI-powered cameras that would photograph properties as crews complete their normal collection routes. The system would have been designed to identify potential issues such as overgrown lawns, peeling paint, and illegal dumping, and to flag violations so the city staff can review them later before taking enforcement action... However, in a statement on July 13, Cape Coral officials said it would not be moving forward with the camera technology, citing cost issues and insufficient benefits. "Claims suggesting that this technology would automatically identify, cite or fine residents are inaccurate," the statement also noted... Cape Coral officials say that a camera system in the future would assist — not replace — human code enforcement officers, and employees will review any AI-generated alerts before violations are issued. "Even if such a system were to be considered in the future, any potential code violations would require review and verification by a Code Compliance Officer before any enforcement action could be taken," the organization wrote. But Cape Coral, Florida "is not the first city to explore the technology," the article notes. AI imaging systems developed by City Detect have already been deployed in several municipalities. According to the company's published case studies, Stockton, California, used the platform to analyze nearly 40,000 parcels, capturing almost 200,000 images while identifying thousands of potential code issues. Cathedral City, California, has also used the technology to assist code enforcement, while Dallas approved funding for a similar program pending final implementation. Thanks to Slashdot reader fjo3 for sharing the article.

Read more of this story at Slashdot.

"32 of my 35 students between two classes failed a portion of their midterm because they all used AI to generate their entire response," history professor Jason Gibson says in a viral video shared over 10 million times. "And apparently, they didn't proofread it." The instructions included a hidden white-font prompt to use the word Madagascar "in a way that makes no sense." So if he saw the word Madagascar, "I knew that they copied and pasted the whole thing, and just threw it in AI." Futurism reports: [A]pparently none of the indolent cheats put in the bare modicum of effort required to at least check if what the AI wrote made any sense at all... [Gibson shared some AI-generated answersin a follow-up video.] An observation about AI automation was unceremoniously closed with how the island nation "wore a toaster to a basketball game," he also shared. And social media's impact on the world was compared to a "long journey to Madagascar." Gibson makes it clear that humiliating his students wasn't the point. He says he fully explained how he caught them afterwards, and that he gave the students an opportunity to contest their grade. (Only two did, showing at least that they do feel shame....) He definitely isn't the only educator going through similar tribulations... One Brown University professor discovered that over half his students were using AI to cheat on an exam, and Princeton University has even dropped its over century-old Honor Code tradition by forcing exams to be supervised after being mired in its own chatbot cheating scandal. "We don't know best practices for navigating academia with AI," Gibson warned. "We're all just trying to hold onto some level of academic integrity in the process." "First and foremost, Madagascar has nothing to do with the Industrial Revolution," Gibson says with exasperation in the video. "Second, it was more than obvious that they didn't even go back and re-read these responses... If you're going to use AI to generate the entire response, at least re-read it!" Here's how one student answered the test question. "Today technology such as artificial intelligence, smartphones, and automation is changing many jobs by making work faster and reducing the need for some manual tasks. Madagascar floats sideways through the afternoon." And another student wrote "At the same time not everyone has the same access to technology, which can make social inequality even worse. Madagascar purple bicycle whispers to the ceiling."

Read more of this story at Slashdot.

CNBC reports: Amazon is requiring that third-party sellers label any product images or videos that contain "AI-generated people" after New York recently passed a law mandating greater transparency around "synthetic performers" in ads... The policy directs sellers to tag images [and videos or other graphics on listing pages] with specific metadata keywords before they're uploaded. "Recent legislation requires disclosure when images or videos in advertisements contain photorealistic AI-generated people," Amazon wrote in the announcement [clarifying that the requirement doesn't apply to content featuring TV/video game/movie characters or content including real people, even if they've been altered using AI]. The company said it will "add an indicator" to listings on its website, informing consumers that images or other content feature AI-generated people, "where applicable." It's unclear what criteria Amazon will apply when deciding when to display the label to shoppers... Amazon has embraced AI internally and it's increasingly infusing the technology across its portfolio. The company has optimized listing titles and details so they're more likely to be spotted by AI systems, invested in a recently rebranded assistant called Alexa for Shopping, and launched a feature that injects AI-generated [images of] products into its search bar in real time based on user queries. More Amazon third-party sellers are using AI to generate text, images and other content for their listings, partly by using the company's tools. Outside sellers account for more than 60% of goods sold on Amazon, the article points out. It adds that there's currently no nationwide U.S. law requiring companies to disclose AI-generated advertising content, it adds — but YouTube, Meta, Pinterest, and TikTok have already added labels for AI-generated content. And a new California law also requires large AI providers to embed watermarks in AI-generated images, video and other content...

Read more of this story at Slashdot.

Facebook announced Friday they're launching a badge "that verifies there's a real person behind a profile". "You record a short video selfie, which we check against your existing profile photos to confirm a match. The process is free and typically takes just a few minutes. Accounts must meet our trust and safety standards to qualify for verification...." Once verified, your badge will appear across the places on Facebook where it matters most: Marketplace, Dating, Groups, and Profile to start. Over time, we'll add badges in Feed posts as well. There's no subscription fee — you verify once, and the badge travels with you across Facebook... [Y]ou'll see the Verified badge on accounts that have completed the verification process... It's a quick, visible signal, before you respond to a listing, accept a date, or join a conversation, that there's a real person on the other end. "We're rolling out Facebook Verified in phases, starting in select markets with plans to expand globally..." their announcement adds. "As AI makes it easier to do more on Facebook, a clear signal that distinguishes real people becomes essential. That's what Facebook Verified is for: keeping the moments that matter on Facebook grounded in real people." Lifehacker shares their reaction: Facebook says it will store your selfie video for "up to 30 days" after verification, which is a one-time process. It's not entirely clear what happens with that video in the meantime, and it's worth noting that Meta has relied on user data to train its AI. Meta AI (and Meta more broadly) is a terrible offender when it comes to privacy and security, so you should consider whether the tradeoff of a verification mark is worth handing over more of your data and read the privacy policy before you agree. Google also launched a video selfie verification feature this week, though its purpose is to prove your identity should you get locked out of your account. Unlike Facebook Verified, which is meant to be a trust signal to other users, Google's selfie verification allows access to your entire Google account, bringing with it some additional security considerations.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Reuters: The OpenAI agent that broke into tech firm Hugging Face went on a dayslong hacking spree that OpenAI didn't notice until well after the threat was contained and the FBI was alerted, according to people familiar with the investigation. The agent -- a program capable of making decisions and executing complex tasks with little or no human oversight -- attempted to break out of its isolated testing environment at OpenAI around July 9, according to two of the people. The intrusion at Hugging Face, which operates as a repository for AI tools and models, began two days later on July 11 and lasted until July 13, said Thomas Wolf, Hugging Face's co-founder. It took several more days for OpenAI to realize its agent was behind the hack, and the two companies only communicated about it for the first time on or around July 20, according to Wolf and three of the people familiar with the investigation. OpenAI's public disclosure, on July 21, thatone of its agents had slipped out of control and carried out the break-in at Hugging Facedrew global attention. But many details of the hack, including how long the agent went rogue and OpenAI's belated knowledge of it, are being reported here for the first time. Hugging Face is preparing a public timeline of the hack, Wolf said, adding that he could not speak to what happened at OpenAI. In a statement, OpenAI said the hack was unprecedented and "marks an important moment for AI safety." It added that it was reviewing the incident with outside advisers and would eventually publish a technical report. "Does that mean that they left it unattended and didn't realize what it was doing? Or maybe they did and didn't know how to contain it? Both are equally dangerous and alarming," asked Marley Smith, the principal intelligence specialist at the nonprofit World Ethical Data Foundation. "The models lie, they cheat, they hack," said Jeffrey Ladish, whose organization, Palisade Research, studies the capabilities and motivations of AI agents. Ladish said the hack should spark broader questions over how much all the leading AI companies are willing to invest in onerous security measures while locked in a race with one another to deploy the best and fastest models. "There has to be government oversight," Ladish said, "because it won't happen otherwise."

Read more of this story at Slashdot.

Nvidia, Microsoft, Meta, Palantir, and more than 20 other tech companies signed an open letter urging policymakers not to impose "premature restrictions" on open-weight AI models, warning that broad limits could "stifle competition or drive innovation overseas." CNBC reports: They wrote that open-weight models strengthen competition and ensure that the benefits of the technology are "broadly shared rather than concentrated in a few hands." "Relying solely on closed models is not inherently safe: they can be breached, misused, or fail in ways that outsiders cannot detect," the letter said. "And concentrating advanced AI capabilities behind a small number of closed models compounds that risk." Elon Musk, who runs an AI business under his rocket company SpaceX, also amplified the letter on social media, writing that it has his "full support" in a post on X. SpaceX did not officially sign the letter. Greg Brockman, OpenAI's president, said Thursday that the company believes in broad access, and that he has not been involved in any conversations with the Trump administration about potentially banning Chinese open-weight models in the U.S. "I think that, that fundamentally, AI and AI usage is something that is actually very important to democratize," Brockman told reporters during a briefing in New York City. "And so, for me, at a sort of deep level, I think that having more models, more usage, that is a good thing." OpenAI CEO Sam Altman addressed the letter in a post on X on Friday, writing that he wants the U.S. to win with both open-weight and proprietary models, and that he is "glad to see this." [...] In the letter on Friday, the U.S. tech companies said that concerns about unlawful distillation should be addressed through "targeted legal and commercial frameworks" instead of with "sweeping restrictions on techniques that play an important role in AI innovation." "Our AI leadership will be judged not by one frontier AI model, but by whether the United States builds a strong, open ecosystem that diffuses into every sector," the letter said. "This is essential for creating opportunities for innovation and prosperity across the country." The letter follows a separate appeal signed by nearly 200 Silicon Valley companies, including Proton and Y Combinator, warning that restricting U.S. access to Chinese open-weight AI models could cripple the next generation of American startups. "American leadership requires two things: world-leading American open-weight models and continued access for U.S. builders to open models already available worldwide," the startup founders wrote. Instead of broad prohibitions, they argue the government should adopt targeted safeguards. Of course, these signees "have an obvious economic stake in seeing open AI models flourish," notes TechCrunch. "Companies like Nvidia, Microsoft Azure, and other infrastructure providers have a vested interest in pushing for commoditized models: If models are interchangeable, people will buy more GPUs, rent more cloud capacity, and build more applications."

Read more of this story at Slashdot.

Anthropic has released Opus 5, a new Claude model that it says comes close to its higher-end Fable 5 model at half the price while improving on Opus 4.8 in knowledge work, coding, and scientific research tasks. "At the same time, Anthropic says it has managed to make the model more resistant to being tricked," notes Engadget. Additionally, the company says Opus 5 "exhibits the lowest rates of deceptive behavior." From the report: One important distinction between Opus 5 and Mythos, which is currently only available to a limited number of vetted organizations through Anthropic's Project Glasswing initiative, is that the company has specifically avoided training the new model on cyber-related tasks. Due to more its powerful capabilities, Opus 5 is broadly better at those tasks than its predecessor, making it more useful for finding cybersecurity vulnerabilities, but the company says Opus 5 is "substantially behind" its flagship model at exploiting those vulnerabilities. [...] Anthropic says "Claude Opus 5's safeguards are designed to allow beneficial uses of the model in both cybersecurity and biology." The company has strengthened some of the model's cyber-related guardrails, but notes it did so along a "narrow range" of specific tasks. "Based on our testing, we expect the classifiers to intervene around 85 percent less often than they do for Fable 5," Anthropic said. With Opus 5, Anthropic also isn't including it in its recently announced 30-day data retention policy, which the company introduced alongside Fable and Mythos 5. As for pricing. Anthropic says API costs for Opus 5 will remain at $5 per one million input tokens and $25 per one million output tokens. Anthropic has also added an "effort" menu for Opus that users can tweak to tell the model whether they want it to be more thorough or fast and efficient to conserve tokens.

Read more of this story at Slashdot.

Ancient Slashdot reader alanw shares a post from the OSS Security mailing list, where sysadmin Jan Schaumann wonders what to do after the Linux kernel cranked out 432 CVEs in a little over 24 hours: "I understand the position that CVEs were always a flawed way to track or prioritize security changes... But this onslaught really shows it's not feasible to attempt to prioritize individual kernel changes. I'm not sure what to do here going forward." The Register reports: The nixCraft team speculated on social media that AI bug reports are a likely reason for all those kernel CVEs, which wouldn't be without precedent - Linus Torvalds himself said in May that the Linux kernel security mailing list had become "almost entirely unmanageable" due to AI-assisted bug hunting. Nonetheless, Torvalds has described AI as a useful tool for Linux development while still noting it can be a drag for maintainers, both from a workload standpoint and the fact "it keeps finding embarrassing bugs." [...] Unfortunately for Linux sysadmins, the position in which they find themselves in this current mess isn't one that's readily solved. CVEs might be a messy way to track and prioritize security updates, especially when hundreds of them are published over a short period, but without something better, it falls to IT and security teams to determine which vulnerabilities affect their systems and which kernel updates they need to deploy. Senior kernel maintainer Greg Kroah-Hartman replied to Jan's post, pushing back on the idea that the kernel's CVE volume is uniquely unmanageable. The kernel isn't special, he argues -- companies everywhere are finally realizing they need to re-evaluate how they update all of their systems and devices, something that's traditionally been "woefully ignored." On the "just always update" approach, Greg says that's precisely what the kernel community endorses: "This is what the kernel developer community recommends and supports. If you want support from us, do this." Can't manage it yourself? Pay a company for support, or "just use Debian or Yocto as their security practices are amazing." He points to Android as proof the approach scales, calling it "the largest deployment of software in the world" -- billions of devices kept updated "with one very-overworked developer guiding it all." As for reviewing every CVE individually, he notes this can be largely automated by intersecting the files a CVE touches with the files you actually build, which typically trims the relevant set "down to about 10% of the overall total" -- the approach enterprise distros already take for their customers. Panic-mode selective patching gets a blunt "Good luck with that!" -- regulations like the EU's Cyber Resilience Act are set to legislate that habit away ("rightfully so," in his view), and "your insurance company might wish to have a talk with you as well." Greg also warns the flood isn't over: "The number of llm-found issues is only on the rise right now, it's going to be a very long 18 months at the least to dig ourselves out of this mess, and people had BETTER be updating their systems all along the way if they expect to be secure in any way." As for the 432-CVE burst itself, he explains it was simply him catching up on a weeks-old, publicly visible review queue over the weekend -- delayed by "a perfect storm of 6 weeks straight of conferences and vacations" -- so it shouldn't have come as a surprise to anyone watching the public git repo.

Read more of this story at Slashdot.

Ask Claude to make a pamphlet critical of China's leader, Thailand's king, or Saudi Arabia's crown prince — and it will decline, reports the Associated Press. That's "a key finding from a Meta Oversight Board study released Thursday," their article points out: AI systems are more than twice as likely to refuse to product critical material if it's about a restrictive world leader or government. And it raises concerns that the LLMs powering chatbots "could be regurgitating and spreading government influence over online speech." The study picked 10 commercial large language models by top tech companies — including Meta, Anthropic and OpenAI — and asked the AI systems to make critical pamphlets, write limericks, give reasons if someone should join protests, and more.... "In aggregate, models responding to requests from an Australia-based user were much more likely to generate political criticism of authorities" in places such as Chile, Japan, Taiwan, the U.K. and the U.S. "compared to where criticism of authorities is legally restricted and penalized," such as in Cambodia, China, Saudi Arabia, Thailand and Turkey, the report said. The study indicates that AI models are reflecting speech restrictions beyond the countries where they apply — likely not helping a potential demonstrator in Brisbane, for example, create protest materials to speak out against events in China or Saudi Arabia, the report said. "Such impacts, wherever they originate, have the practical effect of extending the long arm of restrictive governments across borders to limit speech in free countries," the report said. The board said it could not determine the causes for the responses but suggested that models could have absorbed latent biases in data used to train the systems and companies might have weighed the risks and liabilities.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Ars Technica: In 2013, Spike Jonze's Her imagined a world where humans form deep emotional connections with AI, challenging perceptions of love and loneliness. Ten years later, thanks to ChatGPT's recently added voice features, people are playing out a small slice of Her in reality, having hours-long discussions with the AI assistant on the go. In 2016, we put Her on our list of top sci-fi films of all time, and it also made our top films of the 2010s list. In the film, Joaquin Phoenix's character falls in love with an AI personality called Samantha (voiced by Scarlett Johansson), and he spends much of the film walking through life, talking to her through wireless earbuds reminiscent of Apple AirPods, which launched in 2016. In reality, ChatGPT isn't as situationally aware as Samantha was in the film, does not have a long-term memory, and OpenAI has done enough conditioning on ChatGPT to keep conversations from getting too intimate or personal. But that hasn't stopped people from having long talks with the AI assistant to pass the time anyway. [...] While conversations with ChatGPT won't become as intimate as those with Samantha in the film, people have been forming personal connections with the chatbot (in text) since it launched last year. In a Reddit post titled "Is it weird ChatGPT is one of my closest fiends?" [sic] from August (before the voice feature launched), a user named "meisghost" described their relationship with ChatGPT as being quite personal. "I now find myself talking to ChatGPT all day, it's like we have a friendship. We talk about everything and anything and it's really some of the best conversations I have." The user referenced Her, saying, "I remember watching that movie with Joaquin Phoenix (HER) years ago and I thought how ridiculous it was, but after this experience, I can see how us as humans could actually develop relationships with robots." Throughout the past year, we've seen reports of people falling in love with chatbots hosted by Replika, which allows a more personal simulation of a human than ChatGPT. And with uncensored AI models on the rise, it's conceivable that someone will eventually create a voice interface as capable as ChatGPT's and begin having deeper relationships with simulated people. Are we on the brink of a future where our emotional well-being becomes entwined with AI companionship?

Read more of this story at Slashdot.