Large language models can generate fluent responses, emulate tone, and even follow complex instructions; however, they struggle to retain information across multiple sessions. This limitation becomes more pressing as LLMs are integrated into applications that require long-term engagement, such as personal assistance, health management, and tutoring. In real-life conversations, people recall preferences, infer behaviors, and construct mental maps over time. A person who mentioned their dietary restrictions last week expects those to be taken into account the next time food is discussed. Without mechanisms to store and retrieve such details across conversations, AI agents fail to offer consistency and reliability, undermining user trust. The central challenge with today’s LLMs lies in their inability to persist relevant information beyond the boundaries of a conversation’s context window. These models rely on limited tokens, sometimes as high as 128K or 200K, but when long interactions span days or weeks, even these expanded windows fall short. More critically, the quality of attention degrades over distant tokens, making it harder for models to locate or utilize earlier context effectively. A user may bring up personal details, switch to a completely different topic, and return to the original subject much later. Without a robust memory system, the AI will likely ignore the previously mentioned facts. This creates friction, especially in scenarios where continuity is crucial. The issue is not just forgetting information, but also retrieving the wrong information from irrelevant parts of the conversation history due to token overflow and thematic drift. Several attempts have been made to tackle this memory gap. Some systems rely on retrieval-augmented generation (RAG) techniques, which utilize similarity searches to retrieve relevant text chunks during a conversation. Others employ full-context approaches that simply refeed the entire conversation into the model, which increases latency and token costs. Proprietary memory solutions and open-source alternatives try to improve upon these by storing past exchanges in vector databases or structured formats. However, these methods often lead to inefficiencies, such as retrieving excessive irrelevant information or failing to consolidate updates in a meaningful manner. They also lack effective mechanisms to detect conflicting data or prioritize newer updates, leading to fragmented memories that hinder reliable reasoning. A research team from Mem0.ai developed a new memory-focused system called Mem0. This architecture introduces a dynamic mechanism to extract, consolidate, and retrieve information from conversations as they happen. The design enables the system to selectively identify useful facts from interactions, evaluate their relevance and uniqueness, and integrate them into a memory store that can be consulted in future sessions. The researchers also proposed a graph-enhanced version, Mem0g, which builds upon the base system by structuring information in relational formats. These models were tested using the LOCOMO benchmark and compared against six other categories of memory-enabled systems, including memory-augmented agents, RAG methods with varying configurations, full-context approaches, and both open-source and proprietary tools. Mem0 consistently achieved superior performance across all metrics. The core of the Mem0 system involves two operational stages. In the first phase, the model processes pairs of messages, typically a user’s question and the assistant’s response, along with summaries of recent conversations. A combination of global conversation summaries and the last 10 messages serves as the input for a language model that extracts salient facts. These facts are then analyzed in the second phase, where they are compared with similar existing memories in a vector database. The top 10 most similar memories are retrieved, and a decision mechanism, referred to as a ‘tool call’, determines whether the fact should be added, updated, deleted, or ignored. These decisions are made by the LLM itself rather than a classifier, streamlining memory management and avoiding redundancies. The advanced variant, Mem0g, takes the memory representation a step further. It translates conversation content into a structured graph format, where entities, such as people, cities, or preferences, become nodes, and relationships, such as “lives in” or “prefers,” become edges. Each entity is labeled, embedded, and timestamped, while the relationships form triplets that capture the semantic structure of the dialogue. This format supports more complex reasoning across interconnected facts, allowing the model to trace relational paths across sessions. The conversion process uses LLMs to identify entities, classify them, and build the graph incrementally. For example, if a user discusses travel plans, the system creates nodes for cities, dates, and companions, thereby building a detailed and navigable structure of the conversation. The performance metrics reported by the research team underscore the strength of both models. Mem0 showed a 26% improvement over OpenAI’s system when evaluated using the “LLM-as-a-Judge” metric. Mem0g, with its graph-enhanced design, achieved an additional 2% gain, pushing the total improvement to 28%. In terms of efficiency, Mem0 demonstrated 91% lower p95 latency than full-context methods, and more than 90% savings in token cost. This balance between performance and practicality is significant for production use cases, where response times and computational expenses are critical. The models also handled a wide range of question types, from single-hop factual lookups to multi-hop and open-domain queries, outperforming all other approaches in accuracy across categories. Several Key takeaways from the research on Mem0 include: Mem0 uses a two-step process to extract and manage salient conversation facts, combining recent messages and global summaries to form a contextual prompt.   Mem0g builds memory as a directed graph of entities and relationships, offering superior reasoning over complex information chains.   Mem0 surpassed OpenAI’s memory system with a 26% improvement on LLM-as-a-Judge, while Mem0g added an extra 2% gain, achieving 28% overall. Mem0 achieved a 91% reduction in p95 latency and saved over 90% in token usage compared to full-context approaches.   These architectures maintain fast, cost-efficient performance even when handling multi-session dialogues, making them suitable for deployment in production settings.   The system is ideal for AI assistants in tutoring, healthcare, and enterprise settings where continuity of memory is essential. Check out the Paper. Also, don’t forget to follow us on Twitter and join our Telegram Channel and LinkedIn Group. Don’t Forget to join our 90k+ ML SubReddit. Asif RazzaqWebsite |  + postsBioAsif Razzaq is the CEO of Marktechpost Media Inc.. As a visionary entrepreneur and engineer, Asif is committed to harnessing the potential of Artificial Intelligence for social good. His most recent endeavor is the launch of an Artificial Intelligence Media Platform, Marktechpost, which stands out for its in-depth coverage of machine learning and deep learning news that is both technically sound and easily understandable by a wide audience. The platform boasts of over 2 million monthly views, illustrating its popularity among audiences.Asif Razzaqhttps://www.marktechpost.com/author/6flvq/Multimodal AI on Developer GPUs: Alibaba Releases Qwen2.5-Omni-3B with 50% Lower VRAM Usage and Nearly-7B Model PerformanceAsif Razzaqhttps://www.marktechpost.com/author/6flvq/Diagnosing and Self- Correcting LLM Agent Failures: A Technical Deep Dive into τ-Bench Findings with Atla’s EvalToolboxAsif Razzaqhttps://www.marktechpost.com/author/6flvq/Beyond the Hype: Google’s Practical AI Guide Every Startup Founder Should ReadAsif Razzaqhttps://www.marktechpost.com/author/6flvq/Tutorial on Seamlessly Accessing Any LinkedIn Profile with exa-mcp-server and Claude Desktop Using the Model Context Protocol MCP

The Viwoods AI Paper isn’t your average tablet. Instead of a bright AMOLED or LCD screen, it comes with an eye-friendly e-ink touch display and stylus, turning it into an Android-powered digital notebook. It’s compatible with the Google Play Store, so you can access your favorite apps (like Kindle) and use it for much more, however, and comes with its own built-in AI assistant. It won’t be for everyone, and that’s by design, but if you’re looking for a reading tablet alternative to the Remarkable or Kindle Scribe, it’s a great option. Viwoods AI Paper – PhotosViwoods AI Paper – Design and FeaturesViwoods is a new company, completing its first Kickstarter only late last year. Despite being fresh to the world of consumer electronics, it has debuted with a very strong first product. The Viwoods AI Paper is intuitive, innovative, and surprisingly versatile for an e-ink tablet, and, even while the company is regularly updating it to incorporate new features, it’s already competing with the most popular competitors that have, in some cases, a multi-year head start.The AI Paper comes in two forms, the full-size 10.65-inch version I was sent for testing and the more compact AI Paper Mini, which has an 8.2-inch screen and comes with a backlight the larger version lacks. For this review, I’ll be focusing on the larger version, but if you’re looking for something smaller, those are the most important differences to be aware of. The AI Paper falls into a tablet category that’s better categorized as a Digital Notebook. You’re probably familiar with the Remarkable, which really pushed this category into the mainstream. Digital Notebook tablets don’t claim to be do-everything devices. Instead, they emphasize note-taking, writing, reading, and organization. They’re less prone to cause distraction and, when used properly, can become a pivotal organization and thought-management tool. Like its key competitors in this space, the AI Paper uses a paper-like e-ink display. In this case, the company has implemented the Carta 1300, which offers improved contrast compared to older e-ink screens and, theoretically, should be one of the best out there. Viwoods has done an excellent job of making the “paper” of the screen look bright, but the same can’t be said for most of the alternatives out there, like the Boox Max, which is noticeably darker grey. The screen has a resolution of 2,560 x 1,920, giving it 300 pixels per inch (PPI) of clarity. It’s a high resolution for an e-ink device and allows text and images to appear crisp and readable. Unlike OLED or LCD displays, e-ink works on an entirely different principle. It's best thought of a bit like an Etch-a-Sketch, where the Etch-a-Sketch has a layer of magnetic filings beneath the screen that are then pulled upwards to create its lines. E-ink displays have a layer of black ink beneath the display that is then pulled into place with the application of electricity. Once the image has been formed, it doesn't need to reset, which means that, in theory, if you are reading a book, each page can be presented as a static image, dramatically increasing battery life and reducing eye strain. This quality, in combination with the matte texture used for the screen, works to make the display look much more like a sheet of paper. Amazon popularized this tech with the original Kindle and it has been a hit for e-readers ever since. This is true of all e-ink tablets, but the AI Paper takes its capabilities further with Google Play Store access. It doesn't come with installed by default but allows you to add it as an app, giving you access to a whole ecosystem of apps, including Google Drive, Microsoft 365, Obsidian, and more. This functionality inherently takes away from its distraction-free nature, but the slow refresh rate of its e-ink screen means that it's unappealing to use the device for things like YouTube or social media, so it evens out. AI comes into play prominently, but can also be completely ignored if you choose to. There are three touch buttons on the bottom, similar to a normal Android device, but the right right button is dedicated to summoning your AI agent. When held, you can speak a prompt which is then sent to ChatGPT. This functionality is also built into its different reading and writing applications with quick commands to analyze what's on the page, generate a new article based on it, convert handwriting to text, or just to summon an AI assistant. It also allows you to come up with your own custom commands so you don't need to retype a prompt every time. Other tablets also feature AI assistance, such as those from Boox, which are also based on Android and have access to the Google Play Store. After all, if you can download any Android app, that also means you can download any Android-compatible LLM. What makes this different, however, is that Viwoods has given it the ability to see and interpret anything on your screen versus just answering text prompts. This is especially useful functionality for students. For example, if your professor has assigned you an article to read, you can generate summaries and study aids quickly and easily simply by creating a prompt for it. You can also take handwritten notes in class and then have the AI transcribe these into written text that can then be accessed on a computer. The transcription process isn't perfect, especially if you have messier handwriting, but it's seen improvements since the tablet launched and is usually close enough that I can understand what I was attempting to write anyway. Its traditional notebook and organizational features are exceptional. The homepage is broken into different sections for writing, organization, drawing, and apps. For writing, there are 31 different templates you can choose from, including everything from normal lined paper to graphic organizers to music notation and even body diagrams for clothing design. Templates are essentially backgrounds that you can write on and take the place of having a paper version of the sheet or organizer. When you’re ready to write, you can choose from seven different implements, including a calligraphy pen, pencil, highlight, and a “thinker” which looks like a dry erase marker. The stylus uses the Wacom standard EMR (electro-magnetic resonance) stylus, which is battery-free, has 4096 levels of pressure sensitivity, and is able to detect when the stylus is tilted. This is incredibly useful when drawing as it allows you to shade with the pencil tool. The tablet has a bundle of neat features that go along with this. Each “note” can be tagged for easy organization and searching them up later. Notepads allow you to create layers, similar to a photo editor, so you can add to drawings and note pages without marring anything you’ve written underneath. If you’re drawing a shape, simply holding the stylus in place at the end of your stroke transforms it into a perfect version of that shape with straight lines and curves for circles. And, of course, it can be used as a reader. It natively supports PDFs and EPUB ebooks, but downloading the Kindle app from the Play Store gives you full access to your Kindle Library as well. Reading on e-ink is as close as you can get to paper without it actually being paper, and thanks to its 300 PPI screen, letters are just as crisp as a Kindle Paperwhite. Getting documents onto and off of the device can be done in several different ways. You can upload your documents to the AI Paper management site and they’ll automatically sync to the device the next time it’s online or you can connect it to your PC with a USB Type-C cable and move files on or off like any other mass storage device. You can also email documents or sync them to the cloud. Finally, the AI Paper comes with a 4,100mAh battery that can last around a week when writing two hours a day. This necessitates turning the WiFi off, however, and actually left me a bit disappointed. One of the downsides of running Android and having these additional functions is that they drain the battery significantly faster. A week isn’t bad, but if you’re taking it to several classes a day and then using it for assignments, it won’t even last that long. Viwoods rates it for a month in standby, which does hold true, so if you’re off on vacation or seasonal break, it should still have juice left when you return. Viwoods AI Paper – PerformanceI’ve tested the AI Paper alongside three other leading e-ink tablets (part of an upcoming feature), and all of these devices are after similar things: a paper-like writing experience, fewer distractions, supporting your education and work life, and improving your organization. That side-by-side comparison has highlighted just how much the Viwoods AI Paper has been able to stand out despite the company being so new.For starters, this is the best implementation of the Carta 1300 display yet. One of the most disappointing qualities of it, as implemented in other tablets like the Boox Max, is that the screen is unusually dark. Whites tend to look dark gray, which can impact readability. The Al Paper is the whitest implementation I've seen so far, offering very good contrast and clarity. The writing experience is also great. I wouldn't say it's like paper exactly, but the screen has a minute texture that creates a bit of scratchiness when writing. It feels good if you've never written on an e-ink tablet before; it takes a little getting used to, but I found it much easier than learning to write with a stylus on my Android tablet and its smooth, glassy surface (I eventually gave up on that). The display also uses a soft surface instead of hard glass, which helps it to feel a bit more natural too. The AI Paper was my first e-ink tablet, so I had to go through that learning curve transitioning from paper and typing exclusively. I found it fairly easy to adapt to. It does take a minute to learn its different tools, but it's very smartly designed with intuitive features like lassoing to erase large bits of text or drawing. Sketching on the tablet was also surprisingly good. In fact, it's one of the best. This is largely because of its variety of writing implements. I especially liked the pencil. It just felt so natural to use, much more so than I would have expected any digital writing instrument to feel. You can naturally tilt and angle it for shading with full pressure sensitivity to dial in gradients. This honestly makes a huge difference and works very well with the pressure settings hidden in the tablet's menu system. Between the two, you can really dial in the feeling to make advanced sketches truly possible on this tablet. I like the stylus, though I do think it could feel a bit more premium. It's simple, slim, and plastic. At the same time, it feels like a normal pencil and has a flat section to rest your thumb and keep it oriented correctly. There's a button on this section that can be used for erasing, or you can flip it over and use the backside as an eraser. It's no Lamy, but it feels good to use. It also made me wish other brands started making spring-loaded nibs. If you push hard enough, the nib provides some resistance that almost feels like you're digging into the page. The AI Paper is at its best when you actively make it a daily companion. I'm finding that to be true of all of these tablets, and perhaps it goes without saying. Even so, I've made a point to carry it in situations when I wouldn't normally carry a notebook. Doing so, and using the keyword system, I’ve found that I’m more organized and less forgetful. I work across multiple organizations, and simply using the AI Paper so regularly has inspired me to keep a daily organizer and to-do list. As an adult with ADHD, I’ve come up with a lot of systems to manage, but never stepped into a daily planner with such regularity.Running on Android also has benefits if you prefer to work in apps that don't come preloaded or to use peripherals like a Bluetooth keyboard. This was also something I was highly interested in, as I have been a fan of the Astrohaus Freewrite for quite some time, a device that uses an e-ink screen as part of its attempt to be a modern typewriter. Since I do most of my writing in Google Docs, I downloaded the app, connected a wireless mechanical keyboard, and was off to the races. It worked exceptionally well. Despite feeling quite fully featured, the Viwoods Al Paper is still a work in progress. The company has been releasing updates regularly, adding features like the aforementioned tap-and-hold auto-shapes. It also still needs some features to give it parity with competitors like the Supernote, such as the ability to add multiple keywords to a document for more advanced sorting and connecting ideas. The biggest issue I personally encountered, however, was ghosting. Because of the way the display technology works, all e-ink displays have some form of image retention when changing screens. It's not permanent like it is with OLED gaming monitors, but it can be distracting. While it's not terrible, it is definitely noticeable. More noticeable than I expected. Thankfully, you can easily refresh the screen by dragging down from the top left corner, which wipes these away. You can also set the screen to different refresh speeds which reduces this. That comes with a trade-off, however. You can choose from Best Display, Fast Mode, and Ultra Fast Mode. Each of the two steps up softens the image. So while it’s possible to make the ghosts less visible, it’s also not something I usually wanted to do when a simple swipe would eliminate them entirely.Purchasing GuideThe Viwoods AI Paper has an MSRP of $799 but is most often found between $500 and $600. At the time of writing, it was priced at $589 on Amazon with a $70 instant coupon. It can also be purchased directly from the manufacturer. A smaller version of the tablet is also available, the Viwoods AI Mini and features an 8-inch display as well as a backlight for around $400.

Unprecedented power cuts swept across Portugal, Spain, and parts of France on Monday — instantly unravelling the tech-dependent lives of me and tens of millions of others.  At first, I wasn’t worried. Then the owner of my Lisbon apartment forwarded me a link: an article in The Sun newspaper titled “Spain & Portugal hit by huge power cuts…” I tried to open it, but the page wouldn’t load. After a few minutes of backup power, the mobile networks were dead. I walked down to my local café, hoping to get some more information. “We have no idea what’s going on — nothing is working,” the owner told me, gesturing to her dead till. She was only serving sandwiches and drinks, and like every other shop, could only accept cash. One of the shop’s customers approached me. “They’re saying it might be a Russian cyberattack,” he said, looking remarkably laid back. “Think of it as a free holiday,” he laughed.  I wasn’t laughing. I had no physical cash on me. That meant I couldn’t buy food, and without electricity, I couldn’t cook what I had in the fridge. I also know water networks rely on electric pumps. Plus, I had three young kids at home wondering whether Dad would ever fix the TV.   The 2025 Agenda Trying to keep calm and cool in the 30-degree heat, I left the store searching for an ATM. The first one was completely dead. The next had power, but only let me check my balance. Cash machines need internet to connect to banking servers and authorise withdrawals — without it, they’re just glowing boxes of false hope. “I took the train to Lisbon this morning, but now I can’t get home,” said a man in his 30s who was also trying to withdraw cash to take a taxi. We chatted for a few minutes, speculating on the potential causes of the blackout. “I heard on the radio that it might be something to do with the weather,” he said, referencing a now-debunked claim that the outage was caused by “induced atmospheric vibration” from extreme temperature or pressure shifts disturbing power lines. I wondered whether this was a technical fault or something more malicious.  I walked on in search of more concrete information. Eventually, I found it at the local hospital, the only place I could find with electricity and WiFi, thanks to its backup diesel generators. Scanning the headlines, I was met with some good news. Officials said there was no sign of a cyberattack. Despite all my Hollywood-fuelled fears, Europe wasn’t being invaded and there was definitely no zombie apocalypse.  The bad news, however, was that no one had any idea when power would be restored. Some sources said a few hours. Others said a week, maybe longer. I headed home empty-handed and with few assurances — and no way to fix the TV. As the sun set, I was worried. “Now I get why people build doomsday bunkers,” I said to my wife.  Offline and unprepared The blackout of April 28 was the largest in European history. Over 60 million people were left without power. Mobile networks were knocked out. Card machines and ATMs went dark. Airports closed, trains were left stranded on the tracks, and with the traffic light disabled, congestion piled up. Some people were trapped in elevators and underground metro systems for hours. In Spain, at least five people are thought to have died due to the incident.     In a matter of minutes, modern European life, so dependent on electricity, data, and instant connectivity, ground to a halt. While power was restored to most of Spain and Portugal by midnight on Monday, the incident has raised serious questions about Europe’s infrastructure, security, and resilience. For me, it’s also sparked a more personal reckoning with my tech dependence, and just how unprepared I am for when the systems I rely on go dark. I’ve always seen myself as pretty self-reliant. I grew up in South Africa, where (planned) blackouts were part of life and unpredictability came baked into the day. I used to camp, forage, and cook meals over open fires. I like making things with my hands. I thought I had a good handle on surviving without the grid. But like most, especially in the west, I’ve become deeply tethered to technology. When the blackout hit, all those systems collapsed — and so did the illusion of my independence. My fridge was full, but I couldn’t cook. My phone had a charge, but no signal. My apps were blind. My digital wallet might as well have been Monopoly money.   It wasn’t just the gadgets that failed — it was the mental safety net they’d built. No news, no way to contact family, no idea how big the outage was. I realised I’d outsourced most of my decision-making to invisible infrastructures.  The kids were calm, oblivious to the potential risks. We played board games by candlelight and watched the sun set over a city lit only by headlights and moonlight. But in my head, I was calculating: how long would the tap water run? Did we have enough food that didn’t need cooking? What if this lasted longer? What if it happened again? The blackout only lasted a day. But the aftershocks — at least for me — haven’t stopped. I’m now thinking about storing cash at home, keeping extra water, buying a gas stove, and a radio. Maybe even moving to the countryside where I could grow my own food. The blackout was a wake-up call. Not to prep for the end of the world, but to reckon with just how much I’d handed over to systems I don’t control — and barely understand.  Story by Siôn Geschwindt Siôn is a freelance science and technology reporter, specialising in climate and energy. From nuclear fusion breakthroughs to electric vehic (show all) Siôn is a freelance science and technology reporter, specialising in climate and energy. From nuclear fusion breakthroughs to electric vehicles, he's happiest sourcing a scoop, investigating the impact of emerging technologies, and even putting them to the test. He has five years of journalism experience and holds a dual degree in media and environmental science from the University of Cape Town, South Africa. When he's not writing, you can probably find Siôn out hiking, surfing, playing the drums or catering to his moderate caffeine addiction. You can contact him at: sion.geschwindt [at] protonmail [dot] com Get the TNW newsletter Get the most important tech news in your inbox each week. Also tagged with

Apple has lost its latest legal battle with Epic Games over App Store commissions. Epic CEO Tim Sweeney says Fortnite will return to the U.S. App Store next week. And with Apple set to report quarterly earnings tomorrow, the timing couldn’t be more pointed. So, what happens now? We’ll share Apple’s official response as soon as one exists. In the meantime, three possible paths lie ahead. Option 1: Appeal Apple could appeal to the Ninth Circuit Court of Appeals to overturn the enforcement order. But first, Judge Gonzalez Rogers must formally rule that Apple violated the existing injunction. If that happens, Apple can file a notice of appeal and request a stay to pause enforcement while the appeal is pending. To succeed, Apple would likely argue that its 27% commission on digital goods purchased via in-app links to the web either complies with the injunction or is being misinterpreted. Separately, Apple can ask the district court—or the Ninth Circuit, if denied—to pause enforcement. That would allow Apple to continue operating the App Store as-is, provided it can convince the court that a delay wouldn’t harm the public or Epic. However, with Epic already submitting Fortnite for App Store approval, this strategy would likely involve delaying or blocking that release, which could further escalate the legal fight. Option 2: Negotiate Apple could also pursue a middle-ground approach: negotiating a new implementation framework with Epic and the court. The goal would be to satisfy the injunction while protecting Apple’s business interests—though that may be easier said than done. Option 3: Comply The more straightforward route would be full compliance. That would mean removing or modifying the 27% fee and allowing app links to external payment options without added friction. Compliance would likely head off contempt penalties and improve Apple’s standing with the court. But it could also weaken Apple’s control over its App Store model—both in the U.S. and globally. 9to5Mac’s Take Apple has a track record of exhausting every legal avenue, which suggests an appeal is likely. Still, the company could determine that further appeals are unwinnable and might only make things worse. For now, an appeal seems the most probable next step. Renegotiating App Store rules in a way that satisfies both Epic and the court feels like a long shot, but Apple may pursue it to protect key parts of its business model. Full compliance is possible, though unlikely—unless Apple frames its changes as compatible with the injunction, whether Epic and the court agree or not. Add 9to5Mac to your Google News feed.  FTC: We use income earning auto affiliate links. More.You’re reading 9to5Mac — experts who break news about Apple and its surrounding ecosystem, day after day. Be sure to check out our homepage for all the latest news, and follow 9to5Mac on Twitter, Facebook, and LinkedIn to stay in the loop. Don’t know where to start? Check out our exclusive stories, reviews, how-tos, and subscribe to our YouTube channel

Let's face it: it's never been a particularly good time to be an elderly internet user. Folders can be tough to navigate, Facebook statuses are virtually indistinguishable from the Google search bar, and the mouse pointer is just so darn tiny.Unfortunately, it's not about to get any easier. Though fake profiles have long been used to scam undiscerning internet users, a growing trend of live-video fakes fueled by AI is targeting the young and old alike.Called "realtime deepfakes" — deepfakes referring to spoofed images that are growing nearly impossible to distinguish from real life — scammers are now altering their appearance and voices in live videos. It's all powered by generative AI, the same tech that's stuffing the internet with reams of computer-generated slop.A deep dive by 404 Media reveals the ways fraudsters are using the lifelike tech to fool users. One video they obtained showed a young Black man appearing as a white guy with a grey beard to a woman he was talking to over Facetime. Another showed a man appearing as a younger woman.Once a scammer secures their target's confidence through their lifelike voice and appearance, the actual fraud can begin. 404 notes the users of this tech typically deploy romance scams, Medicare fraud, and photo verification schemes.Another horrifying example called the "grandparent scam" involves criminals impersonating a loved one's voice in order to coerce a payout or con an elderly person into revealing sensitive info, which can then be used to access secure accounts. "I was anxious to get the money out; I'd do anything for my grandchildren," said one Canadian grandmother who'd been targeted for $9,000 in Canadian dollars.It's easy to rag on the elderly, but the growing sophistication of these tools means that young people, too, are highly susceptible. Research in 2021, for instance, found that the younger generations of gen Z, millennials and gen X were 34 percent more likely to have lost money to fraud than people over the age of 60, though they were more likely to have been conned by fake products and crypto schemes than romance scams.The 2024 presidential election also caused mountains of deepfake slop to rain down on social media.Used to spread memes and disinformation alike, deepfake tools allowed posters like Elon Musk to spoof figures such as Kamala Harris denigrating herself. The line with parody can be blurry; one viral example is the infamous Trump/Biden duet, in which the two political rivals sing a tender Mandarin love song to each other. Other examples abound, like one audio deepfake that purported to show a Baltimore area principal making derogatory and racist comments, and a financial worker who forked out $25 million after scammers faked a multi-person Zoom call, including an AI fake of the company's chief financial officer.If celebrities, financial workers, and entire communities can find themselves at the mercy of unscrupulous deepfakers, what chances do our elderly have of resisting deepfake scams?Share This Article

Apr 30, 2025Ravie LakshmananThreat Intelligence / Malware Cybersecurity researchers have shed light on a Russian-speaking cyber espionage group called Nebulous Mantis that has deployed a remote access trojan called RomCom RAT since mid-2022. RomCom "employs advanced evasion techniques, including living-off-the-land (LOTL) tactics and encrypted command and control (C2) communications, while continuously evolving its infrastructure – leveraging bulletproof hosting to maintain persistence and evade detection," Swiss cybersecurity company PRODAFT said in a report shared with The Hacker News. Nebulous Mantis, also tracked by the cybersecurity community under the names CIGAR, Cuba, Storm-0978, Tropical Scorpius, UNC2596, and Void Rabisu, is known to target critical infrastructure, government agencies, political leaders, and NATO-related defense organizations. Attack chains mounted by the group typically involve the use of spear-phishing emails with weaponized document links to distribute RomCom RAT. The domains and command-and-control (C2) servers used in these campaigns have been hosted on bulletproof hosting (BPH) services like LuxHost and Aeza. The infrastructure is managed and procured by a threat actor named LARVA-290. The threat actor is assessed to be active since at least mid-2019, with earlier iterations of the campaign delivering a malware loader codenamed Hancitor. The first-stage RomCom DLL is designed to connect to a C2 server and download additional payloads using the InterPlanetary File System (IPFS) hosted on attacker-controlled domains, execute commands on the infected host, and execute the final-stage C++ malware. The final variant also establishes communications with the C2 server to run commands, as well as download and execute more modules that can steal web browser data. "The threat actor executes tzutil command to identify the system's configured time zone," PRODAFT said. "This system information discovery reveals geographic and operational context that can be used to align attack activities with victim working hours or to evade certain time-based security controls." RomCom, besides manipulating Windows Registry to set up persistence using COM hijacking, is equipped to harvest credentials, perform system reconnaissance, enumerate Active Directory, conduct lateral movement, and collect data of interest, including files, credentials, configuration details, and Microsoft Outlook backups. RomCom variants and victims are managed by means of a dedicated C2 panel, allowing the operators to view device details and issue over 40 commands remotely to carry out a variety of data-gathering tasks. "Nebulous Mantis operates as a sophisticated threat group employing a multi-phase intrusion methodology to gain initial access, execution, persistence, and data exfiltration," the company said. "Throughout the attack lifecycle, Nebulous Mantis exhibits operational discipline in minimizing their footprint, carefully balancing aggressive intelligence collection with stealth requirements, suggesting either state-sponsored backing or professional cybercriminal organization with significant resources." The disclosure comes weeks after PRODAFT exposed a ransomware group named Ruthless Mantis (aka PTI-288) that specializes in double extortion by collaborating with affiliate programs, such as Ragnar Locker, INC Ransom, and others. Led by a threat actor dubbed LARVA-127, the financially motivated threat actor utilizes an array of legitimate and custom tools to facilitate each and every phase of the attack cycle: discovery, persistence, privilege escalation, defense evasion, credential harvesting, lateral movement, and C2 frameworks like Brute Ratel c4 and Ragnar Loader. "Although Ruthless Mantis is composed of highly experienced core members, they also actively integrate newcomers to continually enhance the effectiveness and speed of their operations," it said. "Ruthless Mantis has significantly expanded its arsenal of tools and methods, providing them with state-of-the-art resources to streamline processes and boost operational efficiency." Found this article interesting? Follow us on Twitter  and LinkedIn to read more exclusive content we post. SHARE    

Develop and execute Video Technical Support Team's strategy, managing both strategic and tactical priorities on daily basis. Build a world class support team, organization and culture; one that delivers an exemplary experience and thrives on delighting Apply NowLet's start your dream job Apply now Meet JobCopilot: Your Personal AI Job HunterAutomatically Apply to Remote Customer Support JobsJust set your preferences and Job Copilot will do the rest-finding, filtering, and applying while you focus on what matters. Activate JobCopilot SMART DESIGN View company Jobs posted: 3 Tired of Applying to Jobs Manually?Let JobCopilot do it for you.No more spreadsheets. No more copy-pasting. Just set your preferences and let your Al copilot search, match, and apply to jobs while you sleep.Applies for jobs that actually match your skillsTailors your resume and cover letter automaticallyWorks 24/7-so you don't have to Activate JobCopilot

Can struggling Spurs stay on track for their first trophy since 2008?

April 30, 20255 min readU.S. Energy Secretary Tries to Blame Renewables for Spain’s Power OutageIn response to a power outage in Spain and Portugal, the U.S. Department of Energy’s secretary Chris Wright tried to blame the use of solar and wind energy, though the cause of the blackout is not yet clearBy Benjamin Storrow & E&E News A view of the starry Alhambra sky with no light as a widespread power outage strikes Spain and Portugal in Granada, Spain on April 28, 2025. Alex Camara/Anadolu via Getty ImagesCLIMATEWIRE | Spain and Portugal were still in the dark Monday when U.S. Energy Secretary Chris Wright went on television to blame a widespread power outage on renewable energy.“It's very sad to see what's happened to Portugal and Spain and so many people there. But you know, when you hitch your wagon to the weather, it's just a risky endeavor,” Wright told CNBC.The remark represented a thinly veiled swipe at wind and solar, which were powering almost three-quarters of the Spanish grid at the time it went dark. The comments stood in contrast to those made by the CEO of the Spanish grid operator, who said no “definitive conclusions” for the outage had been reached.On supporting science journalismIf you're enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.But they fit a broader pattern for Wright, a former oil field services executive who has sought to paint wind and solar as costly, unreliable energy sources that threaten the reliability of the electric grid.Renewables could have played a role in an outage that left tens of millions without power, grid experts said. But they cautioned against rushing to conclusions, saying that a series of systematic factors likely were needed for the power systems of two countries to go black inside five seconds Monday.“What I would say is this has the hallmarks of being a very complicated event,” said Eamonn Lannoye, managing director for Europe at the Electric Power Research Institute, which works with the utility industry. “It's not going to be cut and dry.”Grid disasters have become political fodder in recent years, fueling debates over the role of intermittent resources such as wind and solar.In 2021, when a winter storm slammed into Texas, Gov. Greg Abbott (R) was quick to blame wind and solar for rolling blackouts that left 4.5 million without power. An investigation by the Federal Energy Regulatory Commission later concluded the state’s power system was insufficiently winterized to survive such a storm. It pinned much of the blame on the state’s natural gas system, which reported widespread freezes during the event.“I've seen this playbook: Day one, blame renewables. Then the facts come out six months later,” said Michael Webber, a professor who studies the power industry at the University of Texas at Austin. “There's always more to the story.”People wait in line to shop for groceries in a dark shop during a widespread power outage that struck Spain and Portugal around midday on Monday, with the cause still unknown in Lisbon, Portugal on April 28, 2025.Stringer/Anadolu via Getty ImagesWright has often taken aim at renewables during his time in office.In his welcome remarks to Energy Department staffers, he attributed rising power costs in Europe to solar and wind. A month later, at an industry conference in Houston, he said that wherever wind and solar generation increased, power prices have followed.Wright on Monday went on CNBC from Poland, where he had gone to announce a deal to help the country build its first nuclear power plant. He was asked by CNBC anchor Brian Sullivan if the blackout in Spain and Portugal had led to the realization "we're going to need a lot of power from all different types of sources?"Wright did not explicitly mention wind and solar in his response, but he claimed Europe’s percentage of global gross domestic product was falling due to “expensive, unreliable energy."“It’s a choice, but it’s a bad choice,” he said.Andrea Woods, a DOE spokesperson, said Wright was responding to a question about the need to diversify energy supplies. “He was not making an assessment on the cause of the blackout,” she said.Experts say there are challenges with running a power grid with growing amounts of renewable resources. Grid operators need to keep supply and demand in constant balance to maintain the system’s electric frequency and inertia. That's easier to do with traditional resources such as coal, gas and nuclear, which rely on large spinning turbines, said Pratheeksha Ramdas, an analyst at Rystad Energy.Wind and solar facilities can be equipped with rotating turbines that provide those services, but few currently are, she said. Batteries can be used to stabilize frequency, but their deployment in Spain and Portugal is limited.“It is still early to draw definitive lessons, as the full investigation is ongoing. However, the event does highlight some broad challenges — particularly the need for fast and flexible support systems to prevent cascading failures,” Ramdas wrote in an email.The outages in Spain and Portugal began shortly after 12:30 p.m. on Monday, when a widespread generation outage occurred in southwestern Spain, said Eduardo Prieto, CEO of Red Eléctrica, the Spanish grid operator.The system responded to the outage, but was hit with another generation outage 1.5 seconds later, he said. That prompted a disruption in power flows between Spain and France and widespread disconnection of renewable resources across the grid. Within five seconds, the voltage of the entire system went to zero.Asked at a press conference Tuesday if renewables had contributed to the outage, Prieto said it was “premature to make any pronouncements” but noted the grid operator is investigating a widespread generation outage in the southwest.“Given the southwest region I mentioned, it's quite possible the affected generation could be solar, but as I said, without the information, we cannot conclude anything definitively,” he said.Webber, the Texas professor, said he was struggling to understand how a sudden loss of power generation, particularly from a solar facility, could trigger such a widespread outage. Electric grids are designed to withstand sudden losses of generation from large power plants. Traditionally, those standards are designed to withstand a large nuclear power plant tripping offline, which would lead to a large disruption in the frequency and inertia of the electric grid.That “one solar farm going offline should have caused so much trouble, that seems suspicious to me,” said Webber, who previously worked for a large French utility. “Something happened there that I can't fully explain.”EPRI’s Lannoye echoed that assessment. The more likely explanation is that the control systems, which connect power plants to the grid, likely failed at a series of facilities. That was the case in 2019, when a lightning strike hit an offshore wind farm and natural gas plant in the United Kingdom. The control systems on those power plants malfunctioned, precipitating a series of events that led to a power outage affecting 1 million people.But there is not enough information to determine why those systems might have failed in the case of Spain and Portugal, or if such controls were even at fault, he said.“It's too early to cast judgment, and it's certainly not going to seal the fate of one technology or another,” Lannoye said. “If a blackout had sealed the fate of a technology before now, we wouldn't have gas, we wouldn't have coal, we wouldn't have nuclear. We’d be left with no options.”This story also appears in Energywire.Reprinted from E&E News with permission from POLITICO, LLC. Copyright 2025. E&E News provides essential news for energy and environment professionals.

Here's some new Borderlands 4 gameplay to go with that revised release date From PlayStation's latest showcase. Image credit: Gearbox News by Matt Wales News Reporter Published on April 30, 2025 Borderlands 4, as we learned this week, is now arriving earlier than expected - but definitely not because of GTA 6, oh no. And with its revised 12th September release date now out there, developer Gearbox has shared more of its latest looty shooter in a new PlayStation showcase. Weighing in at around 20 minutes, Gearbox's PlayStation-hosted Borderlands 4 "deep dive" video spotlights a couple of new Vault Hunters, some of its new weapons, new traversal mechanics, its new planet, and more - all to whip-up enthusiasm for what creative director Graeme Timmins called "hands-down our best Borderlands ever". Borderlands 4 - which supposedly harkens back to a time when the series was more "grounded", according to Gearbox founder Randy Pitchford - features four new Vault Hunters, and two were the focus of the studio's showcase: Vex the Siren and Rafa the Exo-Soldier. A new Borderlands 4 gameplay video if you don't fancy the 20-minute deep dive.Watch on YouTube Vex can utilise supernatural phase energy and conjure deadly minions, while Rafa comes equipped with an experimental exo-suit capable of "digistructing" an arsenal of weapons. In an accompanying PlayStation Blog post, Gearbox noted players can choose from three signature abilities for their Vault Hunters, and there's talk of unique class Traits, special arguments, capstone abilities, and three branching upgrade paths of passive skills. The team also highlighted some of Borderlands 4 new movement abilities - including gliding, dashing, double-jumping, and point-grappling - as they offered a short introduction to the Kairos, the all-new, seamlessly connected planet where the series' latest action unfolds. Expect new faces, returning characters - including Claptrap, Moxxi, and Zane - plus a personalised Digirunner vehicle players can summon to speed around the place. There's plenty more detail to be found across Gearbox's deep dive video and blog post, so if any of that appeals, remember: Borderlands 4 now launches for PlayStation 5, Xbox Series X/S, and PC on 12th September and not two weeks later. Why? That's "100 percent" down to the studio's "confidence in the game and development trajectory backed by actual tasks and bug find/fix rates", according to Pitchford. And really, honestly, nothing to do with publisher Take-Two wanting to nudge it out the way of its other big game, GTA 6.