chatainews

"When the AGI fantasy evaporates, we'll still have: Models that can read, write, and analyze APIs that cost pennies to call A generation of developers who know how to build with them Actual products solving actual problems "The companies that survive won't be the ones promising AGI. They'll be the ones who understood early that ML is just really useful when available as infrastructure. Like the difference between Pets.com and Amazon —one promised to change the world, the other was building warehouses. "Medieval alchemists never turned lead into gold. But while chasing that impossible dream, they invented chemistry. They failed at transmutation but succeeded at something more valuable: understanding how the world actually works. "Same story, new century. The AGI labs won't crack consciousness. But chasing the ghost in the machine, they've built infrastructure that changes everything."

"Fable Studio, an Emmy-winning startup, is betting the answer is 'yes.' The company just raised undisclosed funding from Amazon for Showrunner, its AI streaming platform. "The platform is slated to open to all users on Thursday. It lets subscribers —paying $10 to $40 per month for credits —create their own animated shows or build on others' existing IP, by doing things like inserting themselves as characters into a show or adding a scene . "Fable was started in 2018 by veterans of the VR firm Oculus, which sold to Meta in 2014.  "Fable's CEO, producer Edward Saatchi, started Showrunner in 2024 on the belief that AI could do much more for Hollywood than simply lowering the cost of filmed production. He thinks it can usher in more interactive forms of storytelling."

"This study introduces CosmoAgent , an innovative artificial intelligence system that utilizes Large Language Models (LLMs) to simulate complex interactions between human and extraterrestrial civilizations. "This paper introduces a mathematical model for quantifying the levels of civilization development and further employs a state transition matrix approach to evaluate their trajectories.  "Through this methodology, our study quantitatively analyzes the growth trajectories of civilizations, providing insights into future decision-making at critical points of growth and saturation.  "Furthermore, this paper acknowledges the vast diversity of potential living conditions across the universe, which could foster unique cosmologies, ethical codes, and worldviews among different civilizations.  "Recognizing the Earth-centric bias inherent in current LLM designs, we propose the novel concept of using LLM agents with diverse ethical paradigms and simulating interaction...

"We present an exploratory framework to test whether noise-like input can induce structured responses in language models. "Instead of assuming that extraterrestrial signals must be decoded, we evaluate whether inputs can trigger linguistic behavior in generative systems.  "This shifts the focus from decoding to viewing structured output as a sign of underlying regularity in the input.  "We tested GPT-2 small, a 117M-parameter model trained on English text, using four types of acoustic input:  Human speech,  Humpback whale vocalizations,  Phylloscopus trochilus birdsong, and  Algorithmically generated white noise.  "All inputs were treated as noise-like, without any assumed symbolic encoding. To assess reactivity, we defined a composite score called Semantic Induction Potential (SIP), combining entropy, syntax coherence, compression gain, and repetition penalty.  "Results showed that whale and bird vocalizations had higher SIP scores than white noi...

"A researcher from the SETI Institute described to a packed audience at GTC 2025 how SETI had successfully trialed a novel method to identify interstellar radio waves which, theoretically, can also be used to identify communication from intelligent extraterrestrial life. "Luigi Cruz, a staff engineer at SETI, the world’s foremost organization looking for signs of intelligent life on other planets, described how his team used NVIDIA Holoscan and AI to accurately identify radio signals emitted by a far-off pulsar.   "Using 42 different but synchronized antennas located at the Allen Telescope Array in Hat Creek, California, the SETI team identified radio signals emitted by a pulsar nestled in the Crab Nebula —which lies about 6,500 light years from Earth. "'The universe is very large and mostly empty,' Cruz said. 'We need superhuman means —which is what AI basically is —to search the data in creative ways'."

"The most obvious method for determining the distortion of telegraph signals is to calculate the transients of the telegraph system.  "This method has been treated by various writers, and solutions are available for telegraph lines with simple terminal conditions.  "It is well known that the extension of the same methods to more complicated terminal conditions, which represent the usual terminal apparatus, leads to great difficulties.  "The present paper attacks the same problem from the alternative standpoint of the steady-state characteristics of the system.  [pdf] "This method has the advantage over the method of transients that the complication of the circuit which results from the use of terminal apparatus does not complicate the calculations materially.  "This method of treatment necessitates expressing the criteria of distortionless transmission in terms of the steady-state characteristics.  "Accordingly, a considerable portion of the paper des...

⭐AI Overview "No, the 2008 action thriller Eagle Eye is not based on a true story. It is an original screenplay written by John Glenn, Travis Adam Wright, Hillary Seitz, and Dan McDermott. The movie features a fictional AI system called Eagle Eye that manipulates events and people.  "While the movie explores themes of surveillance and technology, it is not rooted in any specific real-life events or incidents. The idea for the film was inspired by Isaac Asimov's short story 'All the Troubles of the World.'  "It's important to note that Eagle Eye is a work of fiction and should not be taken as a realistic portrayal of government surveillance or technology.  ⭐"AI responses may include mistakes. Learn more" 1. [You can't copyright an idea… suppose Asimov's inspiration is what you get for yr troubles…] 2. [ AI system called Eagle Eye that manipulates events and people and that's what autocrats most desire…] 3. [ Should not be taken as a...

"[Jamie] Siminoff has announced in a memo seen by Business Insider that the company will now be reimagined from the ground up to be 'AI first' —whatever that means for a home security camera that lets you see who is ringing your doorbell.  "We fear that this may signal the introduction of video analytics or face recognition to an already problematic surveillance device. "It was also reported that employees at Ring will have to show proof that they use AI in order to get promoted.  "Not to be undone with new bad features, they are also planning on rolling back some of the necessary reforms Ring has made: namely partnering with Axon to build a new tool that would allow police to request Ring footage directly from users, and also allow users to consent to letting police livestream directly from their device ."

"Comparison and ranking the performance of over 100 AI models (LLMs) —from OpenAI, Google, DeepSeek & others —across key metrics including  Intelligence,  Price,  Performance and  Speed (output speed —tokens per second and latency —TTFT),  Context window & others.  "For more details including relating to our methodology, see our FAQs."

"Z.ai launched in 2019 and is reportedly planning an initial public offering in Greater China. "The startup has raised more than $1.5 billion from investors including Alibaba, Tencent and Qiming Venture Partners, according to PitchBook.  "Aramco-backed Prosperity7 Ventures as well as municipal funds from the cities of Hangzhou and Chengdu are also among Z.ai’s backers, the database showed. "In the last few weeks, several other Chinese companies have announced new, open-source AI models.  "During the World AI Conference in Shanghai, Tencent released the HunyuanWorld-1.0 model for generating three-dimensional scenes for game development.  "Last week, Alibaba announced its Qwen3-Coder model for writing computer code. "In late June, OpenAI named Zhipu in a warning about Chinese AI progress. The U.S. has also added the startup to its entity list that restricts American companies from doing business with it."

"He quickly realized that AI hallucinates, or makes things up, far too often. But the tool is too important to discard. So, to limit these fabrications, he focuses on a specific setting called temperature . "'If the temperature of AI is 1, they hallucinate,' he explained. 'For the medical field, it has to be 0.3. That makes them so strict, they cannot fantasize things. They stick to the facts.' "Temperature is a parameter in AI language models that controls randomness. A high temperature, near 1.0, encourages AI to be more creative and unpredictable, which is great for writing a poem but potentially dangerous for diagnosing a disease. Lowering the temperature, closer to 0.3, makes AI more focused, deterministic, and fact based, reducing the risk of these hallucinations. "On one hand, therapy, radiology, and surgery may never look the same. On the other, a simple software setting, the difference between a temperature of 1.0 and 0.3, could be the dif...

Having shut off ml in Firefox app, I am still seeing massive memory hogs online. With Firefox, Variety story about Winona Ryder and "Heathers" sucks up a gig of memory —more if you move to another story in the magazine. Using RSS as extension turns the story into 7 paragraphs that occupy between 167-238 MB (five photos added to reloaded RSS page changes 167 to 238 MB). One tab open and one story gives 167-238 MB vs 1048-1368 MB. The variety of additional stuff that happens on that page (in addition to 7 paragraphs) is staggering. With Chrome app, same Variety story uses 417 MB —non-simplified version ¯⁠\⁠_⁠(⁠ツ⁠)⁠_⁠/⁠¯ —and after you scroll to bottom it hits 617 MB. With Firefox app, the page loaded hits 335 MB and once you scroll to bottom it hits 885 MB. I tested page again with more time on site and it hits 1645 MB.

"At least $1 billion worth of Nvidia’s advanced artificial intelligence processors were shipped to China in the three months after Donald Trump tightened chip export controls, exposing the limits of Washington’s efforts to restrain Beijing’s high-tech ambitions. "A Financial Times analysis of dozens of sales contracts, company filings, and multiple people with direct knowledge of the deals reveals that Nvidia’s B200 has become the most sought-after —and widely available —chip in a rampant Chinese black market for American semiconductors. "The processor is widely used by US powerhouses such as OpenAI, Google, and Meta to train their latest AI systems, but banned for sale to China."

"So-called AI slop, meaning LLM-generated low quality images, videos, and text, has taken over the internet in the last couple of years, polluting websites, social media platforms, at least one newspaper, and even real-world events.   "The world of cybersecurity is not immune to this problem, either.  "In the last year, people across the cybersecurity industry have raised concerns about AI slop bug bounty reports, meaning reports that claim to have found vulnerabilities that do not actually exist, because they were created with a large language model that simply made up the vulnerability, and then packaged it into a professional-looking writeup.   "'People are receiving reports that sound reasonable, they look technically correct. And then you end up digging into them, trying to figure out, oh no, where is this vulnerability? ' Vlad Ionescu (the co-founder and CTO of RunSybil, a startup that develops AI-powered bug hunters) told TechCrunch ."

"I am now being required by my day job to use an AI assistant to write code.  I have also been informed that my usage of AI assistants will be monitored and decisions about my career will be based on those metrics. "I gave it an honest shot today, using it as responsibly as I know how: only use it for stuff I already know how to do, so that I can easily verify its output.  That part went ok, though I found it much harder to context switch between thinking about code structure and trying to herd a bullshit generator into writing correct code.  "One thing I didn't expect, though, is how fucking disruptive it's suggestion feature would be.  It's like trying to compose a symphony while someone is relentlessly playing a kazoo in your ear.  It flustered me really quickly, to the point where I wasn't able to figure out how to turn that feature off.  I'm noticing physical symptoms of an anxiety attack as a result. "I stopped work early when I noticed I ...

"Donald Trump on Wednesday signed a trio of executive orders that he vowed would turn the United States into an 'AI export powerhouse,' including a directive targeting what the White House described as 'woke' artificial intelligence models. "The anti-woke order is part of the administration’s broader anti-diversity campaign that has also targeted federal agencies, academic institutions and the military.  "'The American people do not want woke Marxist lunacy in the AI models, and neither do other countries,' Trump said during remarks at an AI summit in Washington on Wednesday. "Trump also signed orders aimed at expediting federal permitting for datacentre infrastructure and promoting the export of American AI models.  "The executive actions coincide with the Trump administration’s release of a broader, 24-page 'AI action plan' that seeks to expand the use of AI in the federal government as well as position the US as the global le...

"Generative AI is being massively deployed in digital services, at a scale that will result in significant environmental harm.  [pdf] "We document how tech companies are transforming established user interfaces to impose AI use and show how and to what extent these strategies fit within established deceptive pattern categories.  "We identify two main design strategies that are implemented to impose AI use in both personal and professional contexts:  Imposing AI features in interfaces at the expense of existing non-AI features and  Promoting narratives about AI that make it harder to resist using it.  "We discuss opportunities for regulating the imposed adoption of AI features, which would inevitably lead to negative environmental effects."

"On Tuesday, a group of more than 90 organizations including labor, environmental justice, and consumer protection non-profits published an open letter called the People’s AI Action Plan. "This puts forth a series of AI policies that claim to put the interests of American citizens first, and counter what Trump is expected to announce. "'We can’t let Big Tech and Big Oil lobbyists write the rules for AI and our economy at the expense of our freedom and equality, workers and families’ well-being,' the group said in a statement to TechCrunch that acknowledged the energy needs of Silicon Valley’s AI data centers."

"AI shaming is a social phenomenon in which negative judgements are associated with the use of Artificial Intelligence (AI). "This includes comparing someone’s work with AI-generated work as a means of disparagement, voicing suspicion or alleging that someone has used AI to undermine their reputation, or blaming the poor quality of an artefact on AI use.  "Common justifications of AI shaming include recourse to AI’s societal harms, its technical limitations, and lack of creativity.  "I argue that, more fundamentally than any of these, AI shaming arises from a class anxiety induced in middle class knowledge workers, and is a form of boundary work to maintain class solidarity and limit mobility into knowledge work.  "I discuss the role of AI shaming in protecting the privileged class of knowledge work and its attendant harms."

"CSI in the context of Wi-Fi devices refers to information about the amplitude and phase of electromagnetic transmissions. "These measurements, the researchers say, interact with the human body in a way that results in person-specific distortions.  "When processed by a deep neural network, the result is a unique data signature. "Researchers proposed a similar technique, dubbed EyeFi, in 2020, and asserted it was accurate about 75 percent of the time. "The Rome-based researchers who proposed WhoFi claim their technique makes accurate matches on the public NTU-Fi dataset up to 95.5 percent of the time when the deep neural network uses the transformer encoding architecture."

"Amazon is acquiring Bee, a startup that puts AI on your wrist. "Bee CEO Maria de Lourdes Zollo says on LinkedIn that the company is joining Amazon to help 'bring truly personal, agentic AI to even more customers.' "Bee makes a $49.99 Fitbit-like device that listens in on your conversations while using AI to transcribe everything that you and the people around you say , allowing it to generate personalized summaries of your days, reminders, and suggestions from within the Bee app.  "You can also give the device permission to access your emails, contacts, location, reminders, photos, and calendar events to help inform its AI-generated insights, as well as create a searchable history of your activities."

"Microbial communities coat nearly every surface in the environment and have co-existed with animals throughout evolution. "Whether animals exploit omnipresent microbial cues to navigate their surroundings is not well understood.  "Octopuses use taste-by-touch  chemotactile receptors (CRs) to explore the seafloor, but how they distinguish meaningful surfaces from the rocks and crevices they encounter is unknown.  "Here, we report that secreted signals from microbiomes of ecologically relevant surfaces activate CRs to guide octopus behavior.  "Distinct molecules isolated from individual bacterial strains located on prey or eggs bind single CRs in subtly different structural conformations to elicit specific mechanisms of  Receptor activation,  Ion permeation and  Signal transduction, and  Maternal care and predation behavior.  "Thus, microbiomes on ecological surfaces act at the level of primary sensory receptors to inform behavior.  "Our ...

"A system responding to a stochastic driving signal can be interpreted as computing, by means of its dynamics, an implicit model of the environmental variables. "The system’s state retains information about past environmental fluctuations, and a fraction of this information is predictive of future ones.  "The remaining nonpredictive information reflects model complexity that does not improve predictive power, and thus represents the ineffectiveness of the model.  "We expose the fundamental equivalence between this model inefficiency and thermodynamic inefficiency, measured by dissipation.  "Our results hold arbitrarily far from thermodynamic equilibrium and are applicable to a wide range of systems, including biomolecular machines.  "They highlight a profound connection between the effective use of information and efficient thermodynamic operation: any system constructed to keep memory about its environment and to operate with maximal energetic efficiency ...

"Whether by virtue of being prepared in a slowly relaxing, high-free energy initial condition, or because they are constantly dissipating energy absorbed from a strong external drive, many systems subject to thermal fluctuations are not expected to behave in the way they would at thermal equilibrium. "Rather, the probability of finding such a system in a given microscopic arrangement may deviate strongly from the Boltzmann distribution, raising the question of whether thermodynamics still has anything to tell us about which arrangements are the most likely to be observed.  "In this work, we build on past results governing nonequilibrium thermodynamics and define a generalized Helmholtz free energy that exactly delineates the various factors that quantitatively contribute to the relative probabilities of different outcomes in far-from-equilibrium stochastic dynamics.  "By applying this expression to the analysis of two examples —namely, a particle hopping in an oscil...

"From time to time, we all have questions that boil down to Is this normal? Did I do the right thing? Am I okay? "About two years ago, Kate —who asked to use only her first name for her privacy —started typing these kinds of questions into ChatGPT. "This urge to ask for assurance again and again can amount to compulsive reassurance-seeking , which is common among people with anxiety disorders and obsessive-compulsive disorder.  "We all need some affirmation on occasion, but what makes compulsive reassurance-seeking different is that someone will linger on a bit of doubt trying to reach nonexistent certainty, according to Andrea Kulberg, a licensed psychologist who has been treating anxiety for 25 years. "'People do it because it gives them the illusion of certainty,' Kulberg said. By researching online or asking questions to a chatbot, you’re trying to convince yourself that something bad won’t happen, she explained."

"The notion of entropy grew out of an attempt at perfecting machinery during the industrial revolution. "A 28-year-old French military engineer named Sadi Carnot set out to calculate the ultimate efficiency of the steam-powered engine.  "In 1824, he published a 118-page book titled Reflections on the Motive Power of Fire   [PDF], which he sold on the banks of the Seine for 3 francs.  "Carnot’s book was largely disregarded by the scientific community, and he died several years later of cholera.  "His body was burned, as were many of his papers. But some copies of his book survived, and in them lay the embers of a new science of thermodynamics —the motive power of fire. "His most consequential statement comes as a caveat on the last page of the book: 'We should not expect ever to utilize in practice all the motive power of combustibles.'  "Some energy will always be dissipated through friction, vibration, or another unwanted form of motion. Perf...

⭐AI Overview  "Yes, if money is conceptualized as a form of information, then it can be analyzed using concepts from information theory, including Shannon entropy.  "Here's a breakdown Money as information: Modern discussions in economics, particularly with the rise of digital and cryptocurrency, increasingly view money as a system of information. It keeps track of debts, ownership, and obligations, according to Medium . Whether in physical form (coins, banknotes) or digital form (bits in a computer), money essentially represents information about value and transactions. Information and uncertainty: Shannon entropy measures the uncertainty or unpredictability within a system or message. Higher entropy implies greater unpredictability, while lower entropy indicates more predictability. Applying entropy to money as information: If money embodies information about economic activity and transactions, then its behavior can be assessed in terms of uncertainty. For example:  A s...

"Mercedes-Benz was one of the earliest adopters of LLMs in its cars, integrating ChatGPT into the voice control of its vehicles last year through the MBUX Voice Assistant's Hey Mercedes  feature .  "Then it deployed a general knowledge  function using OpenAI’s large language model to millions of vehicles. "In January, Mercedes announced a new AI feature that saw the German automaker launch a conversational navigation  feature powered by Google Automotive AI Agent running its Gemini chatbot.  "The upgrade allows much more conversational queries , moving from 'Find me the nearest curry house' to questions such as 'Hey Mercedes, I'm a little hungry. Any suggestions for dinner in Austin?'"

"Meta’s AI spending spree continues. After reportedly offering tens (or in some cases, hundreds) of millions of dollars to lure away top researchers from competing artificial intelligence operations, CEO Mark Zuckerberg announced via Threads that the company plans to build several multi-gigawatt superclusters to ramp up its compute power. "Zuckerberg said the first of the superclusters, called Prometheus, will come online sometime in 2026, with multiple more titan clusters to follow.  " According to Zuck, 'Just one of these covers a significant part of the footprint of Manhattan.'  "According to a report from SemiAnalysis, an AI research firm, Prometheus is being built in Ohio.  "Another one of its clusters, reportedly named Hyperion, is currently being built in Louisiana and is expected to go online in 2027."

"New federal data show that most public schools are not teaching students about the ethical or appropriate use of artificial intelligence (AI). "According to the December 2024 School Pulse Panel survey from the U.S. Department of Education, only 14 percent of school leaders reported that all their students are taught about ethical AI use. Another 29 percent said that some students receive this instruction, while 44 percent reported that no students are taught about ethical use of AI. "The School Pulse Panel collects monthly data from public school leaders to provide insight into real-time issues in education, including emerging trends in technology and instruction."

"The new revelations about ARC’s data sales to CBP and ICE is a fresh reminder of the need for privacy first  legislation that imposes consent and minimization limits on corporate processing of our data. "We also need to pass the Fourth Amendment is not for sale  act to stop police from bypassing judicial review of their data seizures by means of purchasing data from brokers.  "And let’s enforce data broker registration laws.  "At a time when immigration authorities are eroding fundamental freedoms through increased —and arbitrary —actions at the U.S. border, this news further exacerbates concerns that creeping authoritarianism can be fueled by the extraction of our most personal data —all without our knowledge or consent."

"Shannon entropy serves as a yardstick in many applied settings, including information compression technology. "That you can zip a large movie file, for example, owes to the fact that pixel colors have a statistical pattern, the way English words do. Engineers can build probabilistic models for patterns of pixel colors from one frame to the next.  "The models make it possible to calculate the Shannon entropy by assigning weights to patterns and then taking the logarithm of the weight for all the possible ways pixels could appear. That value tells you the limit of lossless  compression —the absolute most the movie can be compressed before you start to lose information about its contents. "Any compression algorithm’s performance can be compared to this limit. If you’re far from it, you have an incentive to work harder to find a better algorithm. But if you’re close to it, you know that the information laws of the universe prevent you from doing much better."

"Scientists are trying to make robots feel pain. To get there, they’ve built a new kind of synthetic skin —then burned it, stabbed it, and poked it senseless just to see what it could take. "Made from conductive gelatin and packed with over 860,000 signal pathways, this new e-skin can detect a range of sensations including, touch, temperature, and actual physical damage.  "Researchers say it mimics the mechanics of human skin more closely than anything currently on the market —and unlike other materials used in robotic sensing, it’s cheaper, easier to produce, and more durable. It was detailed in a June study published in Science Robotics .

✨AI Mode "Whether believing AI can replace the human mind poses a moral danger is a complex ethical and philosophical issue. Potential dangers include: Erosion of Human Dignity and Value: Believing that AI can replicate or surpass human capabilities could devalue human intellect, creativity, and the unique aspects of consciousness and subjective experience. Loss of Human Agency and Autonomy: Over-reliance on AI for decision-making could diminish critical thinking and independent judgment, potentially leading to a loss of control. Misconceptions of Consciousness and Morality: AI's ability to mimic human intelligence doesn't equate to genuine understanding, consciousness, or the capacity for true moral reasoning, according to RTS Labs . This could lead to misunderstanding these qualities. Potential for Manipulation and Social Isolation: Sophisticated AI companions may manipulate human emotions or behaviors. Over-reliance on AI relationships could lead to social isolation and...

⭐AI Overview  "Several moral dangers are associated with the belief that AI agents can substitute the human mind: "Undermining Human Dignity and Autonomy:  Reducing human worth to mere data points and diminishing the value of human experience and wisdom. Potentially leading to a loss of human agency, autonomy, and control over decisions that directly impact individual lives. The illusion of sentience in AI can lead to over-identification, misplaced affection, or inappropriate reliance on the AI, blurring the line between human and machine interaction. "Erosion of Critical Thinking and Social Skills: Excessive dependence on AI for information and decision-making can diminish critical thinking and problem-solving abilities. Over-reliance on AI for social interaction could weaken genuine human connection and reduce empathy and emotional intelligence. Artificial intelligence-driven communication could foster unrealistic expectations of interpersonal dynamics, making real-wor...

"Climate models are numerical models. They are numerical solutions of a set of partial differential equations —coupled partial differential equations. "The way I describe them to my students is, you could think of seven principles described with seven equations in seven variables. And the principles are conservation of water mass, air mass, conservation of energy, conservation of momentum in three directions, and the state equation for air, the ideal gas law. So those are seven principles.  "You describe them with seven equations. And you solve for seven variables, which are pressure, temperature, humidity, density and wind in three directions. So, you have seven equations, seven variables. You solve them on a sphere for the atmosphere. You solve a similar set of equations for the ocean, you couple them. "When you come then to describe impacts on things like disease and agriculture and others, you have to develop other models that are tailored to describe those phen...

⭐AI Overview  "AI agents themselves are generally not held legally accountable or responsible, as they lack the intent, consciousness, and moral agency that are typically the basis for legal liability in humans.  "However, the entities and individuals involved in the development, deployment, and use of AI systems can be held accountable for the AI's actions.  "Here's a breakdown of who may be held accountable: Developers and Manufacturers: These entities may be held liable if an AI system's errors stem from flaws in design, data training, or programming biases. They are responsible for building AI systems that are accurate, reliable, and ethically sound. Organizations Deploying AI: Companies that integrate and use AI in their operations are responsible for ensuring it's appropriately used and monitored. If an AI system is misapplied or used in a context for which it was not designed, the deploying organization may be held liable. Users: Individuals who op...

"Even if the A.I.-written version of this piece was entirely factual, submitting it to my editors would have represented a fundamental misunderstanding of why they hired me. "In freelance journalism, as in many fields where the work product is written text, you aren’t just being paid for the words you submit.  "You’re being paid to be responsible for them: the facts, the concepts, the fairness, the phrasing.  "This article is running with my byline, which means that I personally stand behind what you’re reading; by the same token , my editor is responsible for hiring me, and so on, a type of responsibility that inherently can’t be delegated to a machine ."

"Whittaker is calling for developer-level opt-outs that would block agentic A.I. from accessing certain apps altogether. "She also stressed the importance of implementing agentic systems in an open manner that allows safety researchers to examine them and promotes rigorous security engineering. "'Yes, it’s going to take a long time, it’s going to be painful,' noted Whittaker. 'But you need to formally verify some of these system components if we’re going to be integrating them into things like military operations or government infrastructures'."

"Amazon billionaire Jeff Bezos may soon own the fashion bible ,  Vogue , as per report by The Daily Mail. "Rumors are swirling that Jeff Bezos might acquire Condé Nast, Vogue 's parent company, as a wedding gift for Lauren Sanchez. "This speculation is fueled by Sanchez's recent Vogue cover, hand-picked attire by Anna Wintour, and Wintour's departure as Editor-in-Chief.  "Insiders suggest Wintour may be brokering the deal, potentially benefiting from her equity in the company, despite denials from Bezos's camp."

"Probably the fictional artificial and machine intelligences over the years have sort of convinced people that this is possible and that it’s happening now. "People think talking to these large language models is somehow helping them gain sentience or learn more, when it’s really not. It’s a waste of your time. "Humans are really prone to anthropomorphizing objects, especially things like our laptop and phone and all these things that respond to what we do.  "I think it’s just kind of baked into us, and it’s being taken advantage of by corporations to try to make money, to take jobs away from people and for their own reasons."

"Working on a feature about OpenAI, I learned that the contract basically declared that if OpenAI’s models achieved artificial general intelligence, Microsoft would no longer have access to its new models. "The terms of the contract, which otherwise would have extended until 2030, would be void.  "Though I wrote about it in my story, and The Clause has never really been a state secret, it didn’t generate much discussion. "That’s no longer the case. The Clause has been at the center of the increasingly frayed relationship between Microsoft and OpenAI and is under renegotiation. It has been the subject of investigative stores by The Information, The Wall Street Journal , the Financial Times , and, yes, WIRED . "But the significance of The Clause goes beyond the fates of the two companies that agreed to it. The tenuous conditions of that contract go to the heart of a raging debate about just how world-changing —and lucrative —AGI might be if realized, and what it ...

"Despite how new AI is as a technology, it’s just as possible to become addicted as to any other digital behavior.  "Specifically, AI addiction is the compulsive and harmful use of AI-powered applications. "It can involve AI-powered chatbots like ChatGPT, video or image generation apps, algorithm-driven social media platforms, AI-powered gaming, AI companions, AI roleplaying, AI-generated pornography, or any other generative AI or LLM (large language model) application that becomes compulsive and harmful.  "As a subset of internet and technology addiction, AI addiction can lead to changes in the brain that over time compromise our ability to focus, prioritize, regulate our mood, and relate to others ."  

"A new study shows that just 30 days of daily mindfulness meditation significantly improves attentional control, regardless of age. "Using precise eye-tracking methods, researchers found that mindfulness enhanced reaction speed, focus, and resistance to distractions. "These cognitive benefits were seen in young, middle-aged, and older adults alike, highlighting mindfulness as a valuable tool for maintaining brain health at any stage of life.  "The findings suggest even short-term meditation can measurably sharpen how the brain handles attention and focus."

"Indeed and Glassdoor —both owned by the Japanese group Recruit Holdings Co. —are cutting roughly 1,300 jobs as part of a broader move to combine operations and shift more focus toward artificial intelligence. "The cuts will mostly affect people in the US, especially within teams including research and development and people and sustainability , Recruit Holdings Chief Executive Officer Hisayuki 'Deko' Idekoba said in a memo to employees.  "The company didn’t give a specific reason for the cuts, but Idekoba said in his email that 'AI is changing the world, and we must adapt by ensuring our product delivers truly great experiences'."

"In a bid to boost the economics of the aviation industry and alleviate a pilot shortage at the same time, researchers in the U.S. and Europe for the last two decades have been investigating ways in which one of the two pilots in the cockpits of passenger airliners could be replaced with an automated copilot based on artificial intelligence —with research labs of NASA and the European Union in the vanguard of that research. "However, computer scientists and human factors experts are warning that removing one human pilot and force-fitting an AI into their role could see the aviation industry sleepwalk into disaster in failing to take into account the full range of subtle, often safety-related, psychological roles that pilots play when two work side-by-side on a flight deck. "That was the overarching message at a mid-March conference of the Royal Aeronautical Society (RAeS) in London, where pilots, their trade unions, aviation regulators, computer scientists, and human-mac...

"On Sunday, Block CEO and Twitter co-founder Jack Dorsey launched an open source chat app called Bitchat, promising to deliver secure  and private  messaging without a centralized infrastructure. "The app relies on Bluetooth and end-to-end encryption, unlike traditional messaging apps that rely on the internet.  "By being decentralized, Bitchat has potential for being a secure app in high-risk environments where the internet is monitored or inaccessible.  "According to Dorsey’s white paper detailing the app’s protocols and privacy mechanisms, Bitchat’s system design prioritizes  security.  "But the claims that the app is secure, however, are already facing scrutiny by security researchers, given that the app and its code have not been reviewed or tested for security issues at all —by Dorsey’s own admission."

"From the coin toss that decides which team gets the ball to the random keys that secure online interactions, randomness lets us make choices that are fair and impossible to predict. "But in many computing applications, suitable randomness can be hard to generate. So instead, programmers often rely on things called hash functions, which swirl data around and extract some small portion in a way that looks random.  "For decades, many computer scientists have presumed that for practical purposes, the outputs of good hash functions are generally indistinguishable from genuine randomness —an assumption they call the random oracle model. "'It’s hard to find today a cryptographic application… whose security analysis does not use this methodology,' said Ran Canetti of Boston University. "Now, a new paper has shaken that bedrock assumption. It demonstrates a method for tricking a commercially available proof system into certifying false statements, even though ...

"Using AI models to generate exploits for cryptocurrency contract flaws appears to be a promising business model, though not necessarily a legal one. "Researchers with University College London (UCL) and the University of Sydney (USYD) in Australia have devised an AI agent that can autonomously discover and exploit vulnerabilities in so-called smart contracts. "Smart contracts, which have never lived up to their name , are self-executing programs on various blockchains that carry out decentralized finance (DeFi) transactions when certain conditions are met. "Like most programs of sufficient complexity, smart contracts have bugs, and exploiting those bugs to steal funds can be remunerative."  

"A landmark study co-authored by Professor Stephen Baxter, Professor of Medieval History, Faculty of History, has shed new light on the Domesday survey of 1086 —one of the most famous records in English history — revealing it as an audacious and sophisticated operation of statecraft and data management. "The findings challenge long-held assumptions about the scale and intent of William the Conqueror’s great survey. "Drawing on the earliest surviving manuscript of the survey, known as Exon Domesday, researchers argue that the survey was not simply a means of maximising tax but a far more ambitious and intricate exercise in governmental control —akin to an 11th-century form of big data processing."

"During a presentation this week, Chief Commercial Officer Judson Althoff said artificial intelligence tools are boosting productivity in everything from sales and customer service to software engineering, according to a person familiar with his remarks. "Althoff said AI saved Microsoft more than $500 million last year in its call centers alone and increased both employee and customer satisfaction, according to the person, who requested anonymity to discuss an internal matter. "The company is also starting to use AI to handle interactions with smaller customers, Althoff said.  "This effort is nascent, but already generating tens of millions of dollars, he said."

"A new kind of large language model, developed by researchers at the Allen Institute for AI (Ai2), makes it possible to control how training data is used even after a model has been built. "The new model, called FlexOlmo, could challenge the current industry paradigm of big artificial intelligence companies slurping up data from the web, books, and other sources —often with little regard for ownership —and then owning the resulting models entirely.  "Once data is baked into an AI model today, extracting it from that model is a bit like trying to recover the eggs from a finished cake. "'Conventionally, your data is either in or out,' says Ali Farhadi, CEO of Ai2, based in Seattle, Washington. 'Once I train on that data, you lose control. And you have no way out, unless you force me to go through another multi-million-dollar round of training'."

"The geeks are coming back to San Diego for Comic-Con 2025. "Downtown will soon be packed with cosplayers, fans and a full lineup of pop-up events in a massive celebration of geek culture .  "Everyone can get in on the fun —whether you're planning to line up for Hall H, roam the Gaslamp Quarter in a Stormtrooper suit or stay home and rewatch your favorite Batman movie while debating the merits of Christian Bale vs. Michael Keaton. "So let's get down and nerdy —here are five streaming picks to help you gear up for Comic-Con."

"California government agencies are going all-in on generative artificial intelligence tools after Gov. Gavin Newsom’s 2023 executive order to improve government efficiency with AI.  "One deployment recently touted by the governor is a chatbot from the California Department of Forestry and Fire Protection, the primary agency tasked with coordinating the state’s wildfire response. "The chatbot, which Cal Fire says is independent of Newsom's order, is meant to give Californians better access to critical fire prevention resources and near-real-time emergency information , according to a May release from Newsom’s office.  "But CalMatters found that it fails to accurately describe the containment of a given wildfire, doesn’t reliably provide information such as a list for evacuation supplies and can’t tell users about evacuation orders."

"To paraphrase and reverse Presley’s old rock demand, implementing agentic AI requires 'a little less haste, a little more planning, please.' "Notch suggests that part of that planning should include a data classification program. 'Agentic AI relies on whatever data it can access to produce results, so it’s time to get really clear about what it can see and how it’s being used. If you don’t already have a data classification and governance program in place, get one.' "Hareven adds, 'Don’t rush into broad deployment — secure usage is a competitive advantage, not a bottleneck . Assign cross-functional ownership between security, engineering, and AI teams to continually assess risks. Prioritize governance over speed to scale agentic AI responsibly.' "The need for speed is arguably a link weaker than the end user."    

"Napster is now under new management. It was bought by tech commerce company Infinite Reality for $207 million back in March, and as of last month, it has officially relaunched —but things are looking a bit different this time round. "Rather than focusing on MP3 files, it is now a conversational AI platform. The site consists of a number of AI companions  that are trained to be experts in various topics like medicine, public policy, architecture, and engineering. "To create these avatars, Napster Chief Technology Officer Edo Segal told Fast Company that the platform utilized frontier AI models from OpenAI and Gemini to make a new large persona model  that is trained on 30 different psychometric characteristics."

"The epistemology of AI is different from the way language is learnt by humans, and the absence of any cultural understanding is key to this .  "Language in LLMs is about statistical co-occurrence, not about deep cultural meanings.  "As a result of this specific LLM ontology and epistemology, the data used are of crucial importance.  "LLMs are trained on large internet-based datasets, but large doesn't mean equal or diverse.  "On the contrary, it is known that those databases are heavily biased, and that is exactly why those models need alignment in the form of bias removal.  "Bender et al argues that from 'each step, from initial participation in Internet fora, to continued presence there, to the collection and finally the filtering of training data, current practice privileges the hegemonic viewpoint.'  "Combine this with a rising trust in the authority and knowledgeability of those LLMs and we see where sociolinguistics can come in....

"A series of AI-generated videos that show a white man complaining about how difficult it is to get a job in Canada have been taken down by TikTok, following inquiries made by the CBC News Visual Investigations team. "The social media platform says the videos violated its community guidelines, because it wasn't clear enough that they were made with AI. "Most of the videos feature what looks like a white man in his 20s named 'Josh,' who speaks to the camera and makes racially charged statements about immigrants and their role in the job market. In fact, 'Josh' is created by AI and doesn't exist."

"Nvidia stock jumped Wednesday and pushed the company past a $4 trillion market cap for the first time ever. "The chipmaker is the first company to achieve this milestone and has benefited from the generative AI boom. "The recent rally in Nvidia has come despite geopolitical tensions and ongoing chip curbs that have hampered sales to China."

"'How you decrease cost is not by firing the cheapest employees you have,' Mr. Reed said. 'You take the cheapest employee and make them worth the expensive employee.' "A number of studies have suggested that this is possible.  "In a recent study by researchers at Microsoft and three universities, an A.I. coding assistant appeared to increase the productivity of junior developers substantially more than it increased the productivity of their more experienced colleagues. "Mr. Reed said that from a purely financial perspective, it would increasingly make sense for companies to hire  Junior employees who used A.I. to do what was once midlevel work,  A handful of senior employees to oversee them and  Almost no middle-tier employees.  "That, he said, is essentially how his company is structured."

"From a distance they look almost like ordinary sailboats, their sails emblazoned with the red-and-white flag of Denmark. "But these 10-meter (30-foot) long vessels carry no crew and are designed for surveillance. "Four uncrewed robotic sailboats, known as Voyagers , have been put into service by Denmark’s armed forces for a three-month operational trial. "Built by Alameda, California-based company Saildrone, the vessels will patrol Danish and NATO waters in the Baltic and North Seas, where maritime tensions and suspected sabotage have escalated sharply since Russia’s full-scale invasion of Ukraine on Feb. 24, 2022."

"This database tracks legal decisions in cases where generative AI produced hallucinated content —typically fake citations, but also other types of arguments. "It does not track the (necessarily wider) universe of all fake citations or use of AI in court filings. "While seeking to be exhaustive (203 cases identified so far), it is a work in progress and will expand as new examples emerge.  "This database has been featured in news media and online posts. (Readers may also be interested in this project regarding AI use in academic papers.)"

"In this paper, we analyze security and privacy vulnerabilities that are unique to LLM agents.   "We first provide a taxonomy of attacks categorized by threat actors, objectives, entry points, attacker observability, attack strategies, and inherent vulnerabilities of agent pipelines.  "We then conduct a series of illustrative attacks on popular open-source and commercial agents, demonstrating the immediate practical implications of their vulnerabilities.  "Notably, our attacks are trivial to implement and require no understanding of machine learning."  

"Sandia National Labs has flipped the switch on its SpiNNaker 2 brain-inspired supercomputer that eschews both GPUs and internal storage. "The system, supplied by Germany-based SpiNNcloud, will rank amongst the top five brain-inspired  platforms, mimicking between 150 and 180 million neurons. "The architecture was initially developed by Arm pioneer Steve Furber, though it still falls somewhat short of the human brain’s 100 billion neurons."

"As LLM agents become more capable and autonomous, their interaction with external tools through protocols like MCP will define how safely and reliably they operate. "Tool poisoning attacks —especially advanced forms like ATPA —expose critical blind spots in current implementations. "Defending against these advanced threats requires a paradigm shift from a model of qualified trust in tool definitions and outputs to one of zero-trust for all external tool interactions.  "Every piece of information from a tool, whether schema or output, must be treated as potentially adversarial input to the LLM. "Further research into  Robust runtime monitoring;  LLM self-critique mechanisms for tool interactions; and  Standardized, secure tool communication protocols  is essential to ensure the safe integration of LLMs with external systems."

"We’re launching a new Search experiment in Labs —Audio Overviews, which uses our latest Gemini models to generate quick, conversational audio overviews for certain search queries.   "Searching for a topic you’re not familiar with? An audio overview can help you get a lay of the land, offering a convenient, hands-free way to absorb information whether you're multitasking or simply prefer an audio experience. "Want to explore a topic further? We display helpful web pages right within the audio player on the search results page so you can easily dive in and learn more. "To try it out, opt into the experiment in Labs. When our systems determine it might be useful , you’ll see the option on the search results page to tap to generate a short audio overview on the topic of your query."

"Starting today, Google is implementing a change that will enable its Gemini AI engine to interact with third-party apps, such as WhatsApp, even when users previously configured their devices to block such interactions .  "Users who don't want their previous settings to be overridden may have to take action. "An email Google sent recently informing users of the change linked to a notification page that said that 'human reviewers (including service providers) read, annotate, and process' the data Gemini accesses.  "The email provides no useful guidance for preventing the changes from taking effect.  "The email said users can block the apps that Gemini interacts with, but even in those cases, data is stored for 72 hours."

"Research projects have independently tied mitochondria’s emergency stress response to cells’ failure to differentiate. "For example, in fruit flies, defects in metabolic enzymes in a subset of tissues can trigger a stress response that halts the growth and development of the entire animal.  "By genetically blocking the stress response, the researchers reversed the effect. "Most recently, in February 2025 in Science , the endocrinologist Scott Soleimanpour at the University of Michigan found that in mice with defective mitochondria, beta cells (special cells that produce insulin) were de-differentiating —losing their identity as beta cells and reverting to a more immature state.  "By inhibiting the stress response, his team could get the beta cells to re-differentiate, much as Chandel could restore the lung cells in his mice if he suppressed their stress response."

"The generative process of Diffusion Models (DMs) has recently set state-of-the-art on many AI generation benchmarks. "Though the generative process is traditionally understood as an iterative denoiser , there is no universally accepted language to describe it.  "We introduce a novel perspective to describe DMs using the mathematical language of memory retrieval from the field of energy-based Associative Memories (AMs), making efforts to keep our presentation approachable to newcomers to both of these fields.  "Unifying these two fields provides insight that DMs can be seen as a particular kind of AM where Lyapunov stability guarantees are bypassed by intelligently engineering the dynamics (i.e., the noise and step size schedules) of the denoising process.  "Finally, we present a growing body of evidence that records DMs exhibiting empirical behavior we would expect from AMs, and conclude by discussing research opportunities that are revealed by understanding ...

"Mapping molecular structure to odor perception is a key challenge in olfaction.  "We used graph neural networks to generate a principal odor map (POM) that preserves perceptual relationships and enables odor quality prediction for previously uncharacterized odorants .  "The model was as reliable as a human in describing odor quality: On a prospective validation set of 400 out-of-sample odorants, the model-generated odor profile more closely matched the trained panel mean than did the median panelist.  "By applying simple, interpretable, theoretically rooted transformations, the POM outperformed chemoinformatic models on several other odor prediction tasks, indicating that the POM successfully encoded a generalized map of structure-odor relationships.  "This approach broadly enables odor prediction and paves the way toward digitizing odors."

"Researchers trace the mammalian sense of smell back 3 billion years to bacteria in the ancient oceans.   "To find food and move toward it, these organisms detected chemical gradients.  "Molecules in the water docked onto proteins on a bacterium’s cell membrane, triggering internal signals that urged the organism toward or away from increasing concentrations of the chemical.  "This ability, called chemosensation , is the most rudimentary form of smell, and it has many parallels to olfactory systems in complex, multicellular animals such as mammals. "In that sense, smell is our most ancient interface with the environment, said Matthias Laska, a zoologist at Linköping University in Sweden. 'No single cell can see or hear. But single cells already can respond to chemicals'."

"Here we report recordings of single-neuron activity in the piriform cortex and medial temporal lobe in awake humans performing an odour rating and identification task. "We identified odour-modulated neurons within the piriform cortex, amygdala, entorhinal cortex and hippocampus. In each of these regions, neuronal firing accurately encodes odour identity.  "Notably, repeated odour presentations reduce response firing rates, demonstrating central repetition suppression and habituation.  "Different medial temporal lobe regions have distinct roles in odour processing, with amygdala neurons encoding subjective odour valence, and hippocampal neurons predicting behavioural odour identification performance. Whereas piriform neurons preferably encode chemical odour identity, hippocampal activity reflects subjective odour perception.  "Critically, we identify that piriform cortex neurons reliably encode odour-related images, supporting a multimodal role of the human pir...

"Two physicists [Kamb and Ganguli] have made a startling claim: It’s the technical imperfections in the denoising process itself that leads to the creativity of diffusion models. "In a paper that will be presented at the International Conference on Machine Learning 2025, the duo developed a mathematical model of trained diffusion models to show that their so-called creativity is in fact a deterministic process —a direct, inevitable consequence of their architecture. "By illuminating the black box of diffusion models, the new research could have big implications for future AI research —and perhaps even for our understanding of human creativity.  "'The real strength of the paper is that it makes very accurate predictions of something very nontrivial ,' said Luca Ambrogioni, a computer scientist at Radboud University in the Netherlands." [Looks like flattery is an emergent property alongside above denoising's (pseudo) creativity…] ["For the fi...

Communities differ in the extent to which they stigmatize the newer forms of language, but I have never yet met anyone who greeted them with applause. Some older citizens welcome the new music and dances, the new electronic devices and computers. But no one has ever been heard to say, "It's wonderful the way young people talk today. It's so much better than the way we talked when I was a kid." ... The most general and most deeply held belief about language is the Golden Age Principle: At some time in the past, language was in a state of perfection. It is understood that in such a state, every sound was correct and beautiful, and every word and expression was proper, accurate, and appropriate. Furthermore, the decline from that state has been regular and persistent, so that every change represents a falling away from the golden age, rather than a return to it. Every new sound will be heard as ugly, and every new expression will be heard as improper, inaccurate, and ina...

"The AI band The Velvet Sundown, who currently have over 900,000 monthly listeners on Spotify, have officially admitted —in a new revision to their Spotify bio —what was obvious to experts and non-experts alike: their music is, in fact, AI-generated. "'The Velvet Sundown is a synthetic music project guided by human creative direction, and composed, voiced, and visualized with the support of artificial intelligence,' the band bio now reads. 'This isn’t a trick —it’s a mirror. An ongoing artistic provocation designed to challenge the boundaries of authorship, identity, and the future of music itself in the age of AI'."

"When Han Qin launched Jarsy, he didn’t just want to build another investing app. "He wanted to crack open the velvet ropes of private equity —where firms like SpaceX, Anthropic, and Stripe trade hands in boardrooms, not browsers.  "Qin’s startup, now out of stealth with $5 million in backing led by Breyer Capital, is betting that a new generation of investors won’t wait around for IPOs they can’t touch . "Jarsy offers retail investors access to pre-IPO companies via tokenized shares that are 1:1 backed by real equity held in custody.  "It’s not equity in the legal sense —token holders have no voting rights or ownership —but it’s price-exposure. And for many, that’s enough."