An anonymous reader quotes a report from 404 Media: A "vibe coded" AI app developed by entrepreneur and Y Combinator group partner Tom Blomfield has generated recipes that gave users instruction on how to make "Cyanide Ice Cream," "Thick White Cum Soup," and "Uranium Bomb," using those actual substances as ingredients. Vibe coding, in case you are unfamiliar, is the new practice where people, some with limited coding experience, rapidly develop software with AI assisted coding tools without overthinking how efficient the code is as long as it's functional. This is how Blomfield said he made RecipeNinja.AI. [...] The recipe for Cyanide Ice Cream was still live on RecipeNinja.AI at the time of writing, as are recipes for Platypus Milk Cream Soup, Werewolf Cream Glazing, Cholera-Inspired Chocolate Cake, and other nonsense. Other recipes for things people shouldn't eat have been removed. It also appears that Blomfield has introduced content moderation since users discovered they could generate dangerous or extremely stupid recipes. I wasn't able to generate recipes for asbestos cake, bullet tacos, or glue pizza. I was able to generate a recipe for "very dry tacos," which looks not very good but not dangerous. In a March 20 blog on his personal site, Blomfield explained that he's a startup founder turned investor, and while he has experience with PHP and Ruby on Rails, he has not written a line of code professionally since 2015. "In my day job at Y Combinator, I'm around founders who are building amazing stuff with AI every day and I kept hearing about the advances in tools like Lovable, Cursor and Windsurf," he wrote, referring to AI-assisted coding tools. "I love building stuff and I've always got a list of little apps I want to build if I had more free time." After playing around with them, he wrote, he decided to build RecipeNinja.AI, which can take a prompt as simple as "Lasagna," and generate an image of the finished dish along with a step-by-stape recipe which can use ElevenLabs's AI generated voice to narrate the instruction so the user doesn't have to interact with a device with his tomato sauce-covered fingers. "I was pretty astonished that Windsurf managed to integrate both the OpenAI and Elevenlabs APIs without me doing very much at all," Blomfield wrote. "After we had a couple of problems with the open AI Ruby library, it quickly fell back to a raw ruby HTTP client implementation, but I honestly didn't care. As long as it worked, I didn't really mind if it used 20 lines of code or two lines of code." Having some kind of voice controlled recipe app sounds like a pretty good idea to me, and it's impressive that Blomfield was able to get something up and running so fast given his limited coding experience. But the problem is that he also allowed users to generate their own recipes with seemingly very few guardrails on what kind of recipes are and are not allowed, and that the site kept those results and showed them to other users.

Read more of this story at Slashdot.

An anonymous reader quotes a report from TechCrunch: Anthropic announced on Wednesday that it's launching a new Claude for Education tier, an answer to OpenAI's ChatGPT Edu plan. The new tier is aimed at higher education, and gives students, faculty, and other staff access to Anthropic's AI chatbot, Claude, with a few additional capabilities. One piece of Claude for Education is "Learning Mode," a new feature within Claude Projects to help students develop their own critical thinking skills, rather than simply obtain answers to questions. With Learning Mode enabled, Claude will ask questions to test understanding, highlight fundamental principles behind specific problems, and provide potentially useful templates for research papers, outlines, and study guides. Anthropic says Claude for Education comes with its standard chat interface, as well as "enterprise-grade" security and privacy controls. In a press release shared with TechCrunch ahead of launch, Anthropic said university administrators can use Claude to analyze enrollment trends and automate repetitive email responses to common inquiries. Meanwhile, students can use Claude for Education in their studies, the company suggested, such as working through calculus problems with step-by-step guidance from the AI chatbot. To help universities integrate Claude into their systems, Anthropic says it's partnering with the company Instructure, which offers the popular education software platform Canvas. The AI startup is also teaming up with Internet2, a nonprofit organization that delivers cloud solutions for colleges. Anthropic says that it has already struck "full campus agreements" with Northeastern University, the London School of Economics and Political Science, and Champlain College to make Claude for Education available to all students. Northeastern is a design partner -- Anthropic says it's working with the institution's students, faculty, and staff to build best practices for AI integration, AI-powered education tools, and frameworks. Anthropic hopes to strike more of these contracts, in part through new student ambassador and AI "builder" programs, to capitalize on the growing number of students using AI in their studies.

Read more of this story at Slashdot.

An AI system has for the first time figured out how to collect diamonds in the hugely popular video game Minecraft -- a difficult task requiring multiple steps -- without being shown how to play. Its creators say the system, called Dreamer, is a step towards machines that can generalize knowledge learned in one domain to new situations, a major goal of AI. From a report: "Dreamer marks a significant step towards general AI systems," says Danijar Hafner, a computer scientist at Google DeepMind in San Francisco, California. "It allows AI to understand its physical environment and also to self-improve over time, without a human having to tell it exactly what to do." Hafner and his colleagues describe Dreamer in a study in Nature published on 2 April. In Minecraft, players explore a virtual 3D world containing a variety of terrains, including forests, mountains, deserts and swamps. Players use the world's resources to create objects, such as chests, fences and swords -- and collect items, among the most prized of which are diamonds. Importantly, says Hafner, no two experiences are the same. Every time you play Minecraft, it's a new, randomly generated world," he says. This makes it useful for challenging an AI system that researchers want to be able to generalize from one situation to the next. "You have to really understand what's in front of you; you can't just memorize a specific strategy," he says.

Read more of this story at Slashdot.

Microsoft Chief Technology Officer Kevin Scott has predicted that AI will generate 95% of code within five years. Speaking on the 20VC podcast, Scott said AI would not replace software engineers but transform their role. "It doesn't mean that the AI is doing the software engineering job.... authorship is still going to be human," Scott said. According to Scott, developers will shift from writing code directly to guiding AI through prompts and instructions. "We go from being an input master (programming languages) to a prompt master (AI orchestrator)," he said. Scott said the current AI systems have significant memory limitations, making them "awfully transactional," but predicted improvements within the next year.

Read more of this story at Slashdot.

A new paper [PDF] from the AI Disclosures Project claims OpenAI likely trained its GPT-4o model on paywalled O'Reilly Media books without a licensing agreement. The nonprofit organization, co-founded by O'Reilly Media CEO Tim O'Reilly himself, used a method called DE-COP to detect copyrighted content in language model training data. Researchers analyzed 13,962 paragraph excerpts from 34 O'Reilly books, finding that GPT-4o "recognized" significantly more paywalled content than older models like GPT-3.5 Turbo. The technique, also known as a "membership inference attack," tests whether a model can reliably distinguish human-authored texts from paraphrased versions. "GPT-4o [likely] recognizes, and so has prior knowledge of, many non-public O'Reilly books published prior to its training cutoff date," wrote the co-authors, which include O'Reilly, economist Ilan Strauss, and AI researcher Sruly Rosenblat.

Read more of this story at Slashdot.

An anonymous reader quotes a report from Ars Technica: What does it take to get OpenAI and Anthropic -- two competitors in the AI assistant market -- to get along? Despite a fundamental difference in direction that led Anthropic's founders to quit OpenAI in 2020 and later create the Claude AI assistant, a shared technical hurdle has now brought them together: How to easily connect their AI models to external data sources. The solution comes from Anthropic, which developed and released an open specification called Model Context Protocol (MCP) in November 2024. MCP establishes a royalty-free protocol that allows AI models to connect with outside data sources and services without requiring unique integrations for each service. "Think of MCP as a USB-C port for AI applications," wrote Anthropic in MCP's documentation. The analogy is imperfect, but it represents the idea that, similar to how USB-C unified various cables and ports (with admittedly a debatable level of success), MCP aims to standardize how AI models connect to the infoscape around them. So far, MCP has also garnered interest from multiple tech companies in a rare show of cross-platform collaboration. For example, Microsoft has integrated MCP into its Azure OpenAI service, and as we mentioned above, Anthropic competitor OpenAI is on board. Last week, OpenAI acknowledged MCP in its Agents API documentation, with vocal support from the boss upstairs. "People love MCP and we are excited to add support across our products," wrote OpenAI CEO Sam Altman on X last Wednesday. MCP has also rapidly begun to gain community support in recent months. For example, just browsing this list of over 300 open source servers shared on GitHub reveals growing interest in standardizing AI-to-tool connections. The collection spans diverse domains, including database connectors like PostgreSQL, MySQL, and vector databases; development tools that integrate with Git repositories and code editors; file system access for various storage platforms; knowledge retrieval systems for documents and websites; and specialized tools for finance, health care, and creative applications. Other notable examples include servers that connect AI models to home automation systems, real-time weather data, e-commerce platforms, and music streaming services. Some implementations allow AI assistants to interact with gaming engines, 3D modeling software, and IoT devices.

Read more of this story at Slashdot.

Google's AI arm DeepMind has been holding back the release of its world-renowned research, as it seeks to retain a competitive edge in the race to dominate the burgeoning AI industry. From a report: The group, led by Nobel Prize-winner Sir Demis Hassabis, has introduced a tougher vetting process and more bureaucracy that made it harder to publish studies about its work on AI, according to seven current and former research scientists at Google DeepMind. Three former researchers said the group was most reluctant to share papers that reveal innovations that could be exploited by competitors, or cast Google's own Gemini AI model in a negative light compared with others. The changes represent a significant shift for DeepMind, which has long prided itself on its reputation for releasing groundbreaking papers and as a home for the best scientists building AI. Meanwhile, huge breakthroughs by Google researchers -- such as its 2017 "transformers" paper that provided the architecture behind large language models -- played a central role in creating today's boom in generative AI. Since then, DeepMind has become a central part of its parent company's drive to cash in on the cutting-edge technology, as investors expressed concern that the Big Tech group had ceded its early lead to the likes of ChatGPT maker OpenAI. "I cannot imagine us putting out the transformer papers for general use now," said one current researcher. Among the changes in the company's publication policies is a six-month embargo before "strategic" papers related to generative AI are released. Researchers also often need to convince several staff members of the merits of publication, said two people with knowledge of the matter.

Read more of this story at Slashdot.

Britain's flagship AI agency will slash the number of projects it backs and prioritize work on defense, environment and health as it seeks to respond to technological advances and criticism of its record. From a report: The Alan Turing Institute -- named after the pioneering British computer scientist -- will shut or offload almost a quarter of its 101 current initiatives and is considering job cuts as part of a change programme that led scores of staff to write a letter expressing their loss of confidence in the leadership in December. Jean Innes, appointed chief executive in July 2023, argued that huge advances in AI meant the Turing needed to modernise after being founded as a national data science institute by David Cameron's government a decade ago this month. "The Turing has chalked up some really great achievements," Innes said in an interview. "[But we need] a big strategic shift to a much more focused agenda on a small number of problems that have an impact in the real world." A review last year by UK Research and Innovation, the government funding body, found "a clear need for the governance and leadership structure of the Institute to evolve." It called for a move away from the dominance of universities to a structure more representative of AI in UK.

Read more of this story at Slashdot.

Anthropic said it will start sweeping physical offices for hidden devices as part of a ramped-up security effort as the AI race intensifies. From a report: The company, backed by Amazon and Google, published safety and security updates in a blog post on Monday, and said it also plans to establish an executive risk council and build an in-house security team. Anthropic closed its latest funding round earlier this month at a $61.5 billion valuation, which makes it one of the highest-valued AI startups. In addition to high-growth startups, tech giants including Google, Amazon and Microsoft are racing to announce new products and features. Competition is also coming from China, a risk that became more evident earlier this year when DeepSeek's AI model went viral in the U.S. Anthropic said in the post that it will introduce "physical" safety processes, such as technical surveillance countermeasures -- or the process of finding and identifying surveillance devices that are used to spy on organizations. The sweeps will be conducted "using advanced detection equipment and techniques" and will look for "intruders."

Read more of this story at Slashdot.

Topaz Labs released Photo AI version 3.6.0 with support for "Super Focus v2", which can refocus blurry photos 500% faster. Here are the details:

• Super Focus v2: A second-generation of Super Focus and now our most powerful sharpening and focus recovery model EVER. Powered by the same architecture as Recovery v2, this model is 500% faster than before.
• Super Focus is MEGA FAST in the Cloud: It’s true! This model is even FASTER in the cloud than some of our standard sharpening and upscaling models. This means the credit cost is super low in comparison to before. If you can’t run Super Focus v2 on your machine - no worries, because you’re gonna love this in the cloud.
• Color Correction, Highlights, and Shadows with Adjust Lighting v2: Our second generation AI color and lighting filter is incredible, and now you can pick whether you want color correction on or off when applying the model to your images. You can also adjust highlights and shadows to boot.
• More Improvements are also packed into this release, including a much-requested Zoom to Fit preference, new image formats and extensions (fancy JPG or JPEG? TIF or TIFF? no problem), and of course many other fixes and optimizations under the hood as well.

Reminder - the latest Topaz Labs Gigapixel version 8.3 is still $15 off:

 

Topaz Labs released Gigapixel 8.3 (now $15 off)

The post Topaz Photo AI v3.6.0 released: refocus blurry photos 500% faster appeared first on Photo Rumors.

OpenAI plans to release a new open-weight language model -- its first since GPT-2 -- in the coming months and is seeking community feedback to shape its development. "That's according to a feedback form the company published on its website Monday," reports TechCrunch. "The form, which OpenAI is inviting 'developers, researchers, and [members of] the broader community' to fill out, includes questions like 'What would you like to see in an open-weight model from OpenAI?' and 'What open models have you used in the past?'" From the report: "We're excited to collaborate with developers, researchers, and the broader community to gather inputs and make this model as useful as possible," OpenAI wrote on its website. "If you're interested in joining a feedback session with the OpenAI team, please let us know [in the form] below." OpenAI plans to host developer events to gather feedback and, in the future, demo prototypes of the model. The first will take place in San Francisco within a few weeks, followed by sessions in Europe and Asia-Pacific regions. OpenAI is facing increasing pressure from rivals such as Chinese AI lab DeepSeek, which have adopted an "open" approach to launching models. In contrast to OpenAI's strategy, these "open" competitors make their models available to the AI community for experimentation and, in some cases, commercialization.

Read more of this story at Slashdot.

OpenAI is having another viral moment after releasing Images for ChatGPT last week, with millions of people creating Studio Ghibli-inspired AI art. In a post on X today, CEO Sam Altman said the company has "added one million users in the last hour" alone. A few days prior he begged users to stop generating images because he said "our GPUs are melting."

Read more of this story at Slashdot.

What happens when you ask Copilot to "write a program that can be run on an iPhone 16 to select 15 random photos from the phone, tint them to random colors, and display the photos on the phone"? That's what TouchDevelop did for the long-discontinued Windows Phone in a 2013 Microsoft Research 'SmartSynth' natural language code generation demo. ("Write scripts by tapping on the screen.") Long-time Slashdot reader theodp reports on what happens when, 14 years later, you pose the same question to Copilot: "You'll get lots of code and caveats from Copilot, but nothing that you can execute as is. (Compare that to the functioning 10 lines of code TouchDevelop program). It's a good reminder that just because GenAI can generate code, it doesn't necessarily mean it will generate the least amount of code, the most understandable or appropriate code for the requestor, or code that runs unchanged and produces the desired results. theodp also reminds us that TouchDevelop "was (like BASIC) abandoned by Microsoft..." Interestingly, a Microsoft Research video from CS Education Week 2011 shows enthusiastic Washington high school students participating in an hour-long TouchDevelop coding lesson and demonstrating the apps they created that tapped into music, photos, the Internet, and yes, even their phone's functionality. This shows how lacking iPhone and Android still are today as far as easy programmability-for-the-masses goes. (When asked, Copilot replied that Apple's Shortcuts app wasn't up to the task).

Read more of this story at Slashdot.

The giant financial news site Bloomberg "has been experimenting with using AI to help produce its journalism," reports the New York Times. But "It hasn't always gone smoothly." While Bloomberg announced on January 15 that it would add three AI-generated bullet points at the top of articles as a summary, "The news outlet has had to correct at least three dozen A.I.-generated summaries of articles published this year." (This Wednesday they published a "hallucinated" date for the start of U.S. auto tariffs, and earlier in March claimed president Trump had imposed tariffs on Canada in 2024, while other errors have included incorrect figures and incorrect attribution.) Bloomberg is not alone in trying A.I. — many news outlets are figuring out how best to embrace the new technology and use it in their reporting and editing. The newspaper chain Gannett uses similar A.I.-generated summaries on its articles, and The Washington Post has a tool called "Ask the Post" that generates answers to questions from published Post articles. And problems have popped up elsewhere. Earlier this month, The Los Angeles Times removed its A.I. tool from an opinion article after the technology described the Ku Klux Klan as something other than a racist organization. Bloomberg News said in a statement that it publishes thousands of articles each day, and "currently 99 percent of A.I. summaries meet our editorial standards...." The A.I. summaries are "meant to complement our journalism, not replace it," the statement added.... John Micklethwait, Bloomberg's editor in chief, laid out the thinking about the A.I. summaries in a January 10 essay, which was an excerpt from a lecture he had given at City St. George's, University of London. "Customers like it — they can quickly see what any story is about. Journalists are more suspicious," he wrote. "Reporters worry that people will just read the summary rather than their story." But, he acknowledged, "an A.I. summary is only as good as the story it is based on. And getting the stories is where the humans still matter." A Bloomberg spokeswoman told the Times that the feedback they'd received to the summaries had generally been positive — "and we continue to refine the experience."

Read more of this story at Slashdot.

The New York Times notes that white-collar workers have faced higher unemployment than other groups in the U.S. over the past few years — along with slower wager growth. Some economists wonder if this trend might be irreversible... and partly attributable to AI: After sitting below 4% for more than two years, the overall unemployment rate has topped that threshold since May... "We're seeing a meaningful transition in the way work is done in the white-collar world," said Carl Tannenbaum, the chief economist of Northern Trust. "I tell people a wave is coming...." Thousands of video game workers lost jobs last year and the year before... Unemployment in finance and related industries, while still low, increased by about a quarter from 2022 to 2024, as rising interest rates slowed demand for mortgages and companies sought to become leaner.... Overall, the latest data from the Federal Reserve Bank of New York show that the unemployment rate for college grads has risen 30% since bottoming out in September 2022 (to 2.6% from 2%), versus about 18% for all workers (to 4% from 3.4%). An analysis by Julia Pollak, chief economist of ZipRecruiter, shows that unemployment has been most elevated among those with bachelor's degrees or some college but no degree, while unemployment has been steady or falling at the very top and bottom of the education ladder — for those with advanced degrees or without a high school diploma. Hiring rates have slowed more for jobs requiring a college degree than for other jobs, according to ADP Research, which studies the labor market.... And artificial intelligence could reduce that need further by increasing the automation of white-collar jobs. A recent academic paper found that software developers who used an AI coding assistant improved a key measure of productivity by more than 25% and that the productivity gains appeared to be largest among the least experienced developers. The result suggested that adopting AI could reduce the wage premium enjoyed by more experienced coders, since it would erode their productivity advantages over novices... [A]t least in the near term, many tech executives and their investors appear to see AI as a way to trim their staffing. A software engineer at a large tech company who declined to be named for fear of harming his job prospects said that his team was about half the size it was last year and that he and his co-workers were expected to do roughly the same amount of work by relying on an AI assistant. Overall, the unemployment rate in tech and related industries jumped by more than half from 2022 to 2024, to 4.4% from 2.9%. "Some economists say these trends may be short term in nature and little cause for concern on their own," the article points out (with one economist noting the unemployment rate is still low compared to historical averages). Harvard labor economist Lawrence Katz even suggested the slower wage growth could reflect the discount that these workers accepted in return for being able to work from home. Thanks to Slashdot reader databasecowgirl for sharing the article.

Read more of this story at Slashdot.

Samsung teased its "AI Vision Inside" refrigerators at January's CES tradeshow. (Its internal sensors can now detect 37 different fresh ingredients and 50 processed foods, generating lists for your cellphone or a screen on your refrigerator's door.) But the refrigerators are part of a larger "AI Home" lineup of screen-enabled appliances with advanced AI features, and Engadget got to see them all together this weekend at Samsung's Bespoke AI conference in Seoul, Korea: The centerpiece of the Bespoke line remains Samsung's 4-door French-Door refrigerator, which is now available with two different-sized screens. There's a model with a smaller 9-inch screen that starts at $3,999 or one with a massive 32-inch panel called the Family Hub+ for $4,699. The former is ostensibly designed for people who want something a bit more discreet but still want access to Samsung's smart features, which includes widgets for your calendar, music, weather, various cooking apps and more. Meanwhile, the larger model is for families who aren't afraid of having a small TV in their face every time they open their fridge. You can even play videos from TikTok on it, if that's what you're into.... For cooking, Samsung's matte glass induction cooktops are mostly the same, but its Bespoke 30-inch single ($3,759) and double ($4,649) wall ovens have...you guessed it, more AI. In addition to a 7-inch display, there are also cameras and sensors inside the oven that can recognize up to 80 different recipes to provide optimal cooking times. But if you prefer to go off-script and create something original, Samsung says the oven will give you the option to save the recipe and temperature settings after cooking the same dish five times. And for a more fun application of its tech, the oven's cameras can record videos and create time-lapses of your baked goods for sharing on social media. When it's time to clean up, Samsung's $1,399 Bespoke Auto Open Door Dishwasher has a few tricks of its own. In this case, the washer uses AI (yet again) and sensors to more accurately detect food residue and optimize cleaning cycles... There's also an "AI Jet Ultra Cordless Stick" vacuum cleaner, which "uses AI to better detect what surface its on to more effectively hoover up dirt and debris." Interestingly, in January Samsung's refrigerators also got a mention in iFixit's "Worst of CES" video.

Read more of this story at Slashdot.

An anonymous reader quotes a report from MIT Technology Review: The first clinical trial of a therapy bot that uses generative AI suggests it was as effective as human therapy for participants with depression, anxiety, or risk for developing eating disorders. Even so, it doesn't give a go-ahead to the dozens of companies hyping such technologies while operating in a regulatory gray area. A team led by psychiatric researchers and psychologists at the Geisel School of Medicine at Dartmouth College built the tool, called Therabot, and the results were published on March 27 in the New England Journal of Medicine. Many tech companies are building AI therapy bots to address the mental health care gap, offering more frequent and affordable access than traditional therapy. However, challenges persist: poorly worded bot responses can cause harm, and forming meaningful therapeutic relationships is hard to replicate in software. While many bots rely on general internet data, researchers at Dartmouth developed "Therabot" using custom, evidence-based datasets. Here's what they found: To test the bot, the researchers ran an eight-week clinical trial with 210 participants who had symptoms of depression or generalized anxiety disorder or were at high risk for eating disorders. About half had access to Therabot, and a control group did not. Participants responded to prompts from the AI and initiated conversations, averaging about 10 messages per day. Participants with depression experienced a 51% reduction in symptoms, the best result in the study. Those with anxiety experienced a 31% reduction, and those at risk for eating disorders saw a 19% reduction in concerns about body image and weight. These measurements are based on self-reporting through surveys, a method that's not perfect but remains one of the best tools researchers have. These results ... are about what one finds in randomized control trials of psychotherapy with 16 hours of human-provided treatment, but the Therabot trial accomplished it in about half the time. "I've been working in digital therapeutics for a long time, and I've never seen levels of engagement that are prolonged and sustained at this level," says [Michael Heinz, a research psychiatrist at Dartmouth College and Dartmouth Health and first author of the study].

Read more of this story at Slashdot.