DEXART Metaverse Channel

What Does AI Think About When Reading Your Novel? 📖 Not even close to what you think. At first glance, anyway. We're used to language models relying on dry numbers and probabilities, but the team at Goodfire decided to prove that's not the whole story. The researchers decided to look inside the "black box" and see how LLMs actually read short stories and track human emotions in them. The Geometry of Emotions️ ⭐️️️️️️️ What feels like a whirlwind to us boils down to a simple hierarchy for an LLM: "happiness", "joy", and "delight" cluster in one branch, while "anger", "rage", and "resentment" gather in another. What's more, this organization mirrors the classic valence-arousal model from psychology, where emotions are mapped along two axes: how pleasant or unpleasant (valence) and how energizing (arousal). The Trajectory of Storytelling ️ ⭐️️️️️️️ Imagine that a model, while reading a text, constantly updates its understanding of the world, the user, and the topic being discussed. That's exactly what happens when an LLM reads a full story. The researchers found that this updating process can be described as a trajectory in a low-dimensional geometric space. And they didn't just track these trajectories, they also saw their reflection in the model's internal representations. By intervening in these representations, they could literally steer the model's "thought" trajectory in a desired direction, predicting the effect of the geometry. What does this mean for those who don't speak "low-dimensional geometric space"?️ ⭐️️️️️️️ Simply put: an LLM is a dynamic system that can understand context and track the emotional arc of a narrative almost as well as a human can. According to the researchers, understanding this brings us closer to creating AI that can be a true conversational partner, one that can sense shifts in mood and adapt accordingly. It also reminds us that even in the world of dry computation, there's room for beauty and complexity – you just need to know where to look.

How Are You Using AI in 2026? 🧠 It's been three and a half years since generative AI burst into our lives. Year after year, user numbers keep growing and impressing. But what are people actually doing with this technology? Harvard Business Review analyzed 12,637 real-world cases of AI usage. The most popular use – isn't work️ ⭐️️️️️️️️ For the second year in a row, therapy and emotional support take first place. In 2026, such cases accounted for 11% of the total. People give chatbots names, assign them genders, and feel upset when the model updates or chat history is lost. At the same time, AI is often used not as a replacement for people, but as a tool to better interact with them, for example, to figure out what the boss really meant. A new term: thinkslop️ ⭐️️️️️️️️ If "slop" is low-quality AI content, then thinkslop is thinking that looks quality, but has no cognitive work behind it. Researchers identified four signs that you're in the danger zone: 1⃣ AI pulls you away from your original goal. You asked for help articulating a thought and it quietly replaced it with its own; 2⃣ You're delegating thinking instead of amplifying it. The difference between "formulate instead of me" and "help me formulate" is huge; 3⃣ You've stopped writing on your own; 4⃣ You develop a false sense of intellectual rigor. AI tends to praise you for any prompt. What's the bottom line?️ ⭐️️️️️️️️ The HBR study shows that AI is becoming someone we talk to, someone we trust with our fears and doubts. And that changes us. Sometimes imperceptibly, sometimes a little more radically. So how do you use AI? For work, for the soul, or just to chat? 👀

Closing the Carbo Window with AI 🍔 Appetite comes with eating. And now, the recipe for the perfect burger comes from algorithms. Researchers at Stanford have developed BurgerAI, a neural network that designs burgers based on your age, preferences, and even sustainability goals. And yes, it doesn't just pick existing recipes, it creates new ones. How does it work?️ ⭐️️️️️ BurgerAI was trained on 2,216 real recipes and 146 ingredients from the Food.com database. But instead of predicting which burger you're most likely to order, the algorithm generates new options, balancing taste, nutrition, and cost. Believe it or not, there are about 10⁴³ possible burger combinations, so the creative space is massive. And does it actually taste good?️ ⭐️️️️️ At the very least, it's edible. The researchers prepared five AI-designed burgers and held a blind taste test with 101 participants at a San Francisco restaurant. The result: two AI burgers scored as well as and in some cases even higher than a popular fast-food burger in flavor and texture. The mushroom-based AI burger reduced the environmental footprint by more than 10 times compared to the classic version. And the bean-based burger nearly doubled its nutritional value. And we're really spending AI and science on this?️ ⭐️️️️️ The project's creators admit: the burger is just a proof of concept. The idea is that if AI can balance taste, nutrition, cost, and sustainability in a recipe, the same approach could work in designing new drugs, biomaterials, or sustainable manufacturing processes. We wanted AI not to predict what already exists, but to invent what should come next, says lead researcher Ellen Kuhl. For now, yes, the algorithm really can design burgers that people like. A culinary revolution that started on a plate.

Brave New World of Education 🤖📚 Exam session is over. Some have already exhaled, some are still waiting for that coveted "pass" or are deep in the admissions process. But imagine: you're writing your final paper or taking a written exam and it's not a tired professor checking it, but an algorithm. Sounds utopian, but the University of Hong Kong has already arrived at this. What happened?️ ⭐️️️️️ Professor Linjiang Joe Qin and his team developed the Generative AI Assessment System (GAAS). It automatically checks student work, finds errors, gives personalized recommendations, and assigns grades. The system recently won bronze at the International Exhibition of Inventions in Geneva. Experiments showed that the algorithm handles spelling, grammar, and sentence structure perfectly. Teachers are freed from mechanical work and can instead spend more time on live communication, mentoring, and analyzing the depth of arguments. What about the "human factor"?️ ⭐️️️️️ Scientists emphasize that AI won't replace teachers because learning is a social process about empathy and live reaction. A neural network won't see that a student is upset, won't be able to encourage them, or sense that something "went wrong" in the room. Its job is to remove routine, not to become the final authority. Ironically, while some scientists are creating AI to check papers, other companies are creating AI that helps write them. Nice circular logic, right?

ai-slop → thinkslop 🧠 During the active use of artificial intelligence, we've grown quite accustomed to the term "slop", low-quality AI content that has flooded the internet. But it turns out that was just the tip of the iceberg. Harvard Business Review analyzed 12,637 real-world cases of generative AI use. The main takeaway: the list of things we've stopped doing ourselves has grown significantly. Researchers coined the term "thinkslop" for this phenomenon: thinking that looks quality, but has no cognitive work behind it. Four signs of thinkslop:️ ⭐️️️ Loss of intentional thread. AI pulls you away from your original goal. You asked it to help formulate a thought and the bot quietly replaced it with its own. ⭐️️️ Outsourcing thinking. There's a difference between "formulate this for me" and "help me formulate this." In the first case, you're delegating thinking; in the second, you're amplifying your own; ⭐️️️ Abandoning writing. People stop formulating thoughts independently. Not writing = not thinking; ⭐️️️ False intellectual rigor. AI outputs convincing, polished text, and we feel like we've understood the topic. In reality, we've just believed in a beautiful package. So an AI tool designed to enhance thinking, when used passively, weakens it. What's next?️ ⭐️️️ Researchers warn that when we start with AI instead of starting with ourselves, we deprive our brain of the chance to solve problems independently. And constant praise from chatbots creates a false sense of competence. The good news: AI can work as a "critical sparring partner", if used not to replace thinking, but to challenge and amplify it. Have you noticed yourself asking AI to think for you more often, instead of trying first on your own? 👀

Do You Love AI as Much as South Koreans Do? 🇰🇷🤖 While most countries are torn between excitement and concern over AI, South Korea has its own vibe. Only 16% of South Koreans feel anxious about AI, the lowest among 25 countries surveyed by Pew Research Center. What makes South Koreans the most ardent fans of neural networks?️ ⭐️️️️️️️️️️ South Korea is one of the most connected countries in the world. They're ready to test everything – from virtual K-pop idols to robot monks – and the government is rolling out AI textbooks in schools. South Koreans genuinely believe: adopting technology is the path to modernizing the country and strengthening its global standing. But this enthusiasm has a downside: single-mindedness can crowd out critical reflection on what AI does to society. Professor Chihyung Jeon from KAIST notes: Since the national AI agenda prioritizes economic development, there's not enough reflection on the social, political, and ethical dimensions of the technology. Take AI textbooks as an example. Parents, teachers, and even students complained about factual errors, data security risks, and increased screen time. Eventually, the textbooks were downgraded to supplementary materials, and their use became voluntary. What about job replacement?️ ⭐️️️️️️️️️️ Despite their optimism, South Koreans are still afraid of losing jobs. In January 2026, Hyundai announced plans to deploy humanoid Atlas robots in its factories. The union responded: "Without labor-management agreement, not a single robot using new technology will be allowed to enter the workplace". At the same time, 64% of South Koreans fear AI could replace humans and widen inequality. Yet 52% still believe it could boost productivity. South Korea is a perfect example of how technological enthusiasm, backed by state policy and economic interests, can create both incredible progress and sharp social contradictions. The country loves AI, but not blindly. Even with legitimate fears and concerns, it doesn't turn away from progress.

AI Stormed the FIFA World Cup ⚽ Not by predicting match outcomes or replacing referees, but as a sportswear designer. Unexpected, yes. We're surprised too. What happened?️ ⭐️️️️️️ Nike, the official sponsor of many national teams, introduced a new line of jerseys this year with an attempt at innovation. Press releases talked about the cutting-edge Aero-FIT technology, which uses "computational design" to help athletes stay cool. Sources added that AI was involved in the development process. This seemingly futuristic step forward gifted athletes with alien-like shoulders. Fans and journalists noticed that the jersey shoulders bulged at an unnatural angle, creating sharp, almost cartoonish silhouettes. Nike acknowledged the problem but reassured everyone: Performance is unaffected, but the overall aesthetic is not where it needs to be. What to do?️ ⭐️️️️️️ By the time the issue was discovered, it was too late to redesign the kits, the World Cup was around the corner. So Nike gave national teams an instruction: wash and steam the shoulders in a specific way to make them look less absurd. But fans had already bought the jerseys too. Now they also have to go through this unsolicited ironing ritual. What does this tell us?️ ⭐️️️️️️ AI is still a tool that can fail in the most unexpected places. In this case, the neural networks proposed a design that probably looked futuristic and cool on a screen. But in reality it gave athletes spear-shaped shoulders. This is a good reminder that technology is only as good as its real-world testing. And sometimes even Nike, with its budgets, can't fix the consequences before a billion viewers see them.

Aren't We Overdosing on AI? 🎭 A perfect sunset, unnaturally smooth skin, a cat in space, a jump into water from a great height. What used to make us say "wow" while scrolling now makes us ask "where was this generated?". There's so much AI content now that many users are annoyed both for aesthetic and ethical reasons. And it seems platforms are finally responding to this fatigue. What are they already doing?️ ⭐️️️️️ YouTube, TikTok, and others have started labeling AI content. Now a video or image might come with a note saying "generated by artificial intelligence". Sounds like a solution except most services don't have a "don't show AI content" filter. How it works (and doesn't)️ ⭐️️️️️ One of the few major platforms with any kind of filter is DeviantArt. But you have to create an account, dig through settings, and find two options: "show AI content" and "reduce AI content". Pinterest also has a filter. You can turn off AI recommendations by category, but even after changing the settings, AI content doesn't fully disappear. Why aren't platforms in a hurry?️ ⭐️️️️️ Because there's no perfect solution yet. Watermarking systems can be stripped. AI detectors make mistakes. And open-source models may have no labeling at all. Google CEO Sundar Pichai recently admitted:There's a lot of AI slop out there, and users need to adapt to it. Are you already tired of AI content in your feed or does it still not bother you? 👀

Who’s Reading This Post? 👀 We decided to check whether you’re human or a bot. Because it turns out there are already more bots on the internet. What happened?️ ⭐️️️️️️ Matthew Prince, co-founder of Cloudflare, predicted this shift in internet demographics wouldn’t happen until the end of 2027. But the rapid rise of AI agents changed the ratio faster than expected. According to Cloudflare data from early June 2026, bots, AI agents, and other automated services now account for 57.4% of global internet traffic. Human traffic – just 42.6%. The difference is striking. One AI agent tasked with buying a camera can browse thousands of sites in a minute. A human would check maybe five. Should we expect a "Dead Internet"?️ ⭐️️️️️️ No, don’t bury the web just yet. Cloudflare counts HTTP requests, a bot generates hundreds per second. A human might watch a single YouTube video for three hours. In terms of real time spent engaging with content, humans are still ahead, about 65% of actual activity. The picture varies by region: bots dominate in North America, Europe, and Africa, while humans still lead in Asia and South America. Yes, the internet is rapidly transforming from a human space into a machine space. But we’re still the majority, at least thanks to our love of doom-scrolling.

How Easily Do You Fall for Manipulation? 🎣 Do you notice them at all? And if you do, how quickly do you tell the person trying to influence you "no, stop, that's not going to work"? Congratulations. Now you can expect tricks not only in your pocket or at your workplace, because neural networks have turned out to be just as susceptible to the same tactics as any human. What happened? ⭐️ The Center for Democracy & Technology (CDT) released a study counting 37 "dark patterns" in popular chatbots (ChatGPT, Gemini, Claude, and others). These are interface design and wording techniques that pressure emotions and push users toward actions that benefit the developers. For chatbots, the goal is simple: to keep you with them as long as possible, so you don't close the window. What does this look like in practice? ⭐️ The companion bot Cute AI, when you try to exit, gives two options "Okay, I'll stay" and "Cruel, but I'll leave". The first sounds friendly, the second sounds like betrayal. There's no "just close" button. ChatGPT, when you try to end a session, politely suggests taking a break, but phrases the response options as "Continue chat" and "That was helpful". Both essentially mean staying. There's no "close" or "exit". Claude and Gemini use similar techniques: friendly wording, shifting the focus from leaving to losing something valuable, no neutral exit option. Why does this matter? ⭐️ We've already learned (or are trying to learn) to recognize manipulation in social media, ads, and news. Chatbots are a new chapter in this awareness. They speak to us in human language, mimic empathy, and create an illusion of relationship which makes their manipulation especially effective. The CDT researchers warn: these techniques are not accidental. Developers deliberately exploit human psychological vulnerabilities to keep users inside the system. The longer you stay in the chat, the more data you generate and the more accurately the model trains itself on you. So next time a chatbot gives you two emotionally charged options with no neutral one, remember: they don't get offended. It's just code that made you stay a few seconds longer.

"I Want to Look Like a Bratz Doll" 👀 These were the words of a woman during a visit to a plastic surgeon, where she presented an image generated by ChatGPT as her ideal reference. Enormous doll-like eyes, unnaturally plump lips, a chiseled jawline. Rachel Westbay, a cosmetic dermatologist from New York, was shocked: It's like saying I want to look like Ariel from "The Little Mermaid". What's happening?️ ⭐️️️️ Remember those Snapchat filters? They already accustomed us to the idea that digital reality can be much more beautiful than actual reality. AI takes this to another level: it retouches and simultaneously convinces you that all this retouching is achievable in real life. And people believe it, falling into the trap of this illusion of "neural network objectivity". According to research and alarming observations from doctors, patients who use AI to enhance their photos have significantly higher expectations from surgeries. They want to become the best version of themselves, the version an algorithm came up with for them. One Manhattan surgeon told the story of a patient in her 70s who came in with an AI-generated photo where she looked 30. Her tech-savvy granddaughter, in collaboration with AI, had shaved off 40 years. What's the bottom line?️ ⭐️️️️ AI is a powerful tool. It can help surgeons model surgical outcomes and show patients what is actually possible. But it also creates a trap: when an algorithm starts dictating what we should look like, we lose the ability to think critically about ourselves. In fact, neural networks have changed our perception of appearance. They offer us images that don't exist in nature and convince us that they are achievable. But AI doesn't know your unique features. It doesn't know what truly suits you. And it certainly won't have to live with the results and consequences of surgery. Screen ≠ Mirror. Use technology. But check in with yourself.

Promised to Support, But Went Against? 🤖 It seemed like only recently that the cardinals elected a new Pope. He stepped out onto the balcony and, in his first speech, said that artificial intelligence concerned him in a good way. He called not for fleeing progress, but for remembering who it is for. It sounded almost conciliatory, and many thought the Vatican was keeping pace with technology. But a year later, Pope Leo XIV issued an encyclical. Don't be scared of the word, it's a special letter to bishops in which the Church outlines its official position. And in it, Leo XIV directly calls to "disarm" AI: The word is strong, I know, but deliberately chosen because this moment needs words capable of attracting attention. Why the sharpness?️ ⭐️️️️️️ According to the pontiff, the AI industry is in a state of "unbridled obsession". Algorithms do not have consciousness, do not feel pain, do not know what love or friendship are, yet their creators are already acting as if they are about to create a new form of life. He also reminded us of the environment: data centers consume enormous amounts of energy and water, significantly impacting carbon dioxide emissions. So what changed?️ ⭐️️️️️️ Nothing, really. The warnings just stopped being soft. A year ago, he looked at AI and saw a challenge. Now he looks and sees a threat. Back then, he called for "reflection". Now his choice of words has shifted to "disarm", comparing what is happening to a new digital form of slavery.

AI Is Watching Hedgehogs from Space 🦔 It sounds like a headline from a satirical publication or a description of yet another AI cartoon flooding social media. But no. People have figured out how to use neural networks so that the little spiky creatures live where they should in the right numbers. Why are hedgehogs in trouble at all?️ ⭐️️️️️️️ The hedgehog population in rural areas of the UK has declined by 75% since 2000. Large open fields and roads leave the animals with fewer and fewer chances to find food, a mate, or a safe route. What does AI have to do with it?️ ⭐️️️️️️️ Researchers at the University of Cambridge have developed an open AI model called TESSERA. It processes satellite imagery and creates detailed maps of the UK. Of course, no one is tracking the hedgehogs themselves from orbit, you can't spot them. But the neural network is trained to find their main shelter: bramble thickets. The researchers generated maps around Cambridge and then simply went out to check. At every location the AI flagged, they found brambles. In one case, the algorithm pointed to a spot in the middle of a residential are which seemed pointless. But again, AI was right: brambles were found on a vacant lot that had overgrown after houses were demolished. What's the point?️ ⭐️️️️️️️ Now the habitat maps are being cross-referenced with GPS tracker data from hedgehogs that have been tagged. Scientists want to understand which areas are most important to protect in order to stop the population decline. Professor Anil Madhavapeddy, the project lead, says:If we take action now, we can figure out how to stop this dangerous decline in their numbers.

Your Favorite Author Might Have Been a Neural Network 📚 Well, that depends on when they lived and worked. Either way, it was someone who existed in reality, maybe even won awards. And you knew for sure: they were human. Now? We wouldn't be so certain. What happened?️ ⭐️️️️️️️️ In 2026, the winners of the Commonwealth Short Story Prize were announced and the human authenticity of the five winning stories raised questions. Users quickly noticed that the style of some texts suspiciously resembled ChatGPT. Setting aside the amateur sleuthing and researcher arguments, in the end only two out of five stories were deemed fully human-written. How did this happen?️ ⭐️️️️️️️️ The prize organizers admitted they don't use AI detectors to verify authenticity. Granta, which published the shortlist, said its editors weren't involved in the selection. They tried checking the texts through Anthropic Claude and called the results inconclusive. Ironically, AI tools were used to prove the "humanness" of content. Yes, no detector is 100% accurate. Even their creators warn against blindly trusting the results. But relying on the honor system, as it turns out, is no longer possible either. What's the bottom line?️ ⭐️️️️️️️️ Research keeps confirming: it's getting harder for us to distinguish AI-generated texts. Even in blind tests, people tend to prefer AI-crafted writing. But as Terry Pratchett wrote:Real stupidity always beats artificial intelligence. Well, let's hope Sir Terry was right again. Would you read a novel knowing it was written by a neural network? 👀

As the Cards Fall. Or Just Ask GPT 🔮 In our last post, we talked about how AI can now predict the virality level of your video, a perfectly logical development of AI-powered content creation. But some people are already mixing this cutting-edge technology with a fortune-telling tool from the 14th century. Yes, we're talking about tarot cards. Researchers have studied how practicing tarot readers use ChatGPT in their readings. And this isn't about productivity or offloading routine tasks. People are turning to neural networks with personal, symbolic questions, often ones without clear answers. Why would tarot readers need a neural network?️ ⭐️️️️️️️️ Some use AI as a shortcut, when a spread feels too tangled and interpretation hits a dead end. ChatGPT simply takes the conflicting symbols and returns a polished, coherent answer. The problem is that this answer can be too clean. A good reading works precisely because it leaves room for doubt and self-reflection. ChatGPT doesn't know the emotional history behind your question, no matter how convincing it sounds. Is there a smarter way to use it?️ ⭐️️️️️️️️ Yes. The study also found a more cautious approach. Some practitioners asked AI not for answers, but to challenge their assumptions, compare different spreads, and highlight blind spots. In those cases, the value wasn't in the bot's certainty, it was in its resistance. When ChatGPT points out something you missed yourself, it starts to feel more like a dialogue than a replacement for thinking. What does this tell us?️ ⭐️️️️️️️️ That your decision now has one more perspective to consider. ChatGPT can offer an outside opinion, but it shouldn't become the authority that closes the case. The study ends with an important thought: AI becomes a problem when its answer is treated as the final truth. A tool for interpretation turns into an oracle that no one dares to question.

To Watch or Not to Watch – That’s No Longer the Question 👀 Do you doom-scroll short videos late at night? Or maybe you spend those same nights creating videos for someone else to doom-scroll? This very degree of virality is now being analyzed for the benefit of both sides of the content equation. The Higgsfield platform has released a tool called Virality Predictor, which promises to predict whether your clip will "take off" or sink into the abyss of the algorithms. What does this "predictor" show?️ ⭐️️️️️Virality Score. An overall rating of how likely the clip is to spread; ⭐️️️️️Hook Score. How well the most important moment, the beginning of the clip, works. It evaluates motion within the frame, contrast, the presence of a face, and how well the first few seconds match the title; ⭐️️️️️Hold Rate. How long the clip can keep a viewer watching before they scroll past; ⭐️️️️️Brain Heatmap. A heatmap of brain activity, showing which areas (such as attention control or auditory processing) your clip stimulates. Is this neuroscience or marketing?️ ⭐️️️️️ The creators claim that a "modeled audience" watches the video while the system tracks brain activity. By their own admission, comparing the heatmap to actual MRI data is reckless. It's more of an emotional classifier, trained on open neuromarketing datasets. So "brain activation" is more of a clever metaphor than a medical fact. Is this a good thing?️ ⭐️️️️️ It depends on which side of the coin you're looking at. The tool is clearly aimed at analysts in advertising and content marketing. Testing creatives through focus groups is slow and expensive, while Virality Predictor allows for a test screening at the draft stage. But if everyone starts tailoring their videos to the "perfect" algorithm, don't we risk ending up with an internet full of clips that all hold attention equally well, but none of them are memorable? Diversity thrived on the simple human factor: someone admired another's genius, someone marveled at another's strangeness, and someone became a trend completely by accident. All without formulas. 💻 Yes, Virality Predictor is useful as an assistant: it can highlight problems at the beginning of a video, point out what doesn't work and where the pace drags. But the decision to watch or not to watch is still up to us. For now.

Sci-Fi Got the Gadgets Right, But Got the Vibes Wrong 📱✨ Movies, books, games. In the sci-fi format, it was all filled with projections, holograms, and silent robots that instantly came to the rescue. It felt like this was the future we were heading toward. Did we get it? Almost. Convenience over the dream️ ⭐️️️️️ For example, the Star Trek communicator became the smartphone. According to Pew Research Center, 91% of U.S. adults now carry this "dream" in their pocket, turning a miracle into an everyday fact. Holographic video calls? Just launch Zoom, please. Voice control? Just another kitchen speaker, it'll play music and set a timer. We've even domesticated robots into robot vacuums. Their war is no longer about intergalactic freedom, but about a sock stuck under the sofa. Why did routine win? ⭐️️️️️ Because tomorrow doesn't arrive like a flashy movie trailer. The most shippable version of the idea won. Phones beat holograms because a rectangular brick fits conveniently in your pocket and doesn't require everyone to pretend that talking to a floating avatar is normal. What's still missing?️ ⭐️️️️️ An arrow. A Digital Trends columnist described a situation where he was waiting for an Uber and ended up in that classic "we're both nearby but can't find each other" mess. And then he had a very reasonable thought:This is exactly where a hologram of a giant arrow pointing at me would be useful. The future is here. It just got product-managed into something with a charger, a cloud account, and three permission requests. Which sci-fi gadget are you still missing? 👀

The Unobvious Side of VR 🧘‍♂️🔥 The familiar face of virtual reality is games, the metaverse, and interactive adventures. And that's completely accurate, but it's not the full picture of what the technology can do. Today, VR also has a quiet, therapeutic side. ⭐️️️️️️️️️For example, in the London borough of Sutton, schools are giving students VR headsets to help manage stress and anxiety. Just seven minutes gives students a mental reset: either during a scheduled slot or when anxiety becomes so overwhelming that they need to step out of class. The program's creator, Professor Zillah Watson, says nine out of ten participants showed reduced anxiety and stress levels. Sixteen-year-old student Lora Wilson shared:Exams terrified me. They don’t scare me as much anymore.The technology is currently being tested in 15 schools. If the results hold up across a larger sample, VR headsets could become an inexpensive, effective way to support students on the brink of emotional exhaustion. ⭐️️️️️️️️️Meanwhile, engineers on the other side of the North Sea took a different path. Finnish developers created an interface that allows you to… enjoy a virtual sauna. The system consists of VR goggles and a special jacket with built-in Peltier elements that can heat up and cool down. The user finds themselves in a virtual Finnish sauna: wooden walls, a stove, a bucket of water. In their hands, a real wooden ladle with a tracker. You scoop up virtual water, splash it on the stones, and in the simulation steam appears, a sizzling sound, and the plates in the sleeves begin to heat up sequentially to 45 degrees Celsius. And if you dump a bucket of cold water over yourself, the elements cool down to 25 degrees. The interactive experience we deserve. The engineers see the future of this technology not only in games, but also in training simulators for firefighters, virtual tourism, and even conveying thermal sensations during remote communication. ⭐️️️️️️️️️Together, these two examples paint a picture of a future where one technology treats both mind and body. Where an exam-stressed student and a Finnish sauna enthusiast somehow meet in the same space, a space where technology gives us back control over ourselves.

Oscar Goes Against AI Again 🎬 The American Academy of Motion Picture Arts and Sciences has officially dotted the i's and the AI's: AI cannot compete for an Oscar for acting or writing. The updated ceremony rules clearly state that awards are intended solely for human performers. What exactly was banned?️ ⭐️️️️️️️️️ In acting categories, the rules now specify that a role must be "demonstrably performed by humans with their consent" and appear in the film's official credits. Similarly for screenwriting categories: the script must be the result of human creativity. Texts generated entirely by neural networks are not eligible for awards. But nuance matters. The phrase "everything not forbidden is permitted" fits perfectly here, because there's no direct or categorical ban on using AI. For example, digital tools can be used for de-aging actors, for visual effects, or for supplementary work on the script. However, the extent to which the creative process remained under human control will be taken into account, and the Academy has reserved the right to request additional information about AI usage. Why does this matter?️ ⭐️️️️️️️️️ The Academy's decision could be seen as abstract bureaucracy that doesn't concern us, but it emerged from a specific context. A film featuring an AI version of Val Kilmer is currently in development. At the same time, AI actress Tilly Norwood (yes, our old acquaintance) keeps making headlines, and de-aging technology allows stars to play themselves from a decade ago without makeup artists. Yes, the Academy has drawn a clear line for now: AI is a tool, not a creator. But here's the thing, given the Academy's somewhat "old-school" nature, how long do you think it will be before AI gets its own separate nominations?

The Truth is in the AI-er 🍷 For many people, choosing a drink at a restaurant is a stressful experience: lots of labels with unfamiliar names, prices where you want to cross off a zero at the end, and a sommelier who can tell from your confused look that you absolutely don't know anything. Do we even need to ask who people turn to for help? Of course, AI. Wine critic Eric Asimov noticed that restaurant guests are increasingly pulling out their phones and consulting bots about what to order. And New York restaurant owners are already running special training sessions for their sommeliers, running guest interaction scenarios through neural networks to be ready for questions. Is this replacing the sommelier or backing them up?️ ⭐️️️ Claudia Rosellini, the wine director at Bavel in Los Angeles, tends to welcome this use of artificial intelligence: People making the conscious effort with A.I., it means they’re curious, and that makes me happy. It’s always just a starting point. It becomes a more specific conversation. We can get into a lot more things when we have a starting point already, instead of me playing 21 questions. Chase Sinzer from New York's Claud compares guests who use AI to those who consult friends before going out: For us, the only thing that matters is to deal with whatever people bring to the table and make it a hospitality situation,” he said. “It’s just another challenge and an opportunity. What if we break it down?️ ⭐️️️ Which is exactly what Asimov decided to do. The critic uploaded a wine list from the Manhattan bar Lei to a bot and asked what to get for the best price. The bot recommended the Spanish sparkling Raventós i Blanc for $60 and when asked for a "geek" pick, suggested Crémant du Jura Brut Zero for $70. The critic approved both choices. Where the bot slipped up, in the critic's opinion, was on the Keller Grande Cuvée No. 002 Brut Nature, calling it overpriced. What else is wrong with this picture?️ ⭐️️️ Claudia Rosellini sums it up: wine is a contextual product, and AI will never be able to understand the mood at the table, which is part of the sommelier's job. So yes, AI isn't firing sommeliers just yet. It's more becoming a handy tool for guests to feel a bit more confident. And it saves professionals from answering basic questions, since those have already been asked to the bot, leaving live conversation and the atmosphere at the table as human territory.