ResidentialBusiness

European governments are under pressure to lift military spending View the full article

London-based group is testing its self-driving cars in Germany and the US and hopes to ‘soon’ agree deals with automakersView the full article

This article is republished with permission from Wonder Tools, a newsletter that helps you discover the most useful sites and apps. Subscribe here. Typing isn’t always the best way to get your thoughts down. Sometimes talking through an idea leads to better clarity. New AI tools can reliably transform those spoken thoughts into clean, organized text. I’ve spent months experimenting with voice AI tools—first on my phone, and now on my laptop. They’ve been helping me pull ideas from my brain onto paper. The tools below have become crucial to my workflow. Why voice AI beats traditional transcription Traditional transcription simply converts speech to text. Modern voice AI does much more: Instant transformation: Speak naturally and get a polished draft, outline, or summary Smart cleanup: AI removes filler words and adds proper punctuation Format flexibility: Convert speech into various formats like bullet lists or structured documents Context awareness: AI understands context and organizes your thoughts logically. Because it’s grounded in your own words, it doesn’t hallucinate. 5 ways I like using voice AI Here are some scenarios where voice AI is particularly valuable: 1. Journal entries Instead of staring at a blank page, I speak my thoughts at day’s end. The AI transforms my stream of consciousness into organized reflections. 2. Meeting follow-ups After an in-person meeting, I open my voice AI app, hit record, and talk through key points while they’re still fresh. I don’t worry about the structure of my sentences or about pausing as I think. The AI waits for me and summarizes my rambling. 3. Presentation planning Speaking through presentation ideas helps me figure out my narrative flow. The AI helps me organize my thoughts into a structured outline. I can talk through multiple potential versions, then compare them on screen later. 4. Book notes To preserve insights from something I’m reading, I turn on a voice AI app and flip through the pages or scroll through the text to remind myself out loud about intriguing passages or ideas. I then save the structured note the AI creates. I like being able to look back at the text while dictating the note. And the editing part of my brain interferes less when I’m talking than when I’m typing. 5. Daily planning Starting my day by verbally mapping out my priorities helps me think through what’s ahead more effectively than typing out a list. Voice AI apps to try Letterly Easy to use: Just press the app’s big button. Up to 15 minutes per recording. Cross-platform: Record or access your past text-from-voice across automatically synchronized desktop, web, and mobile apps. Smart format detection: The magic transform option can automatically reformat your words, turning lists into bullets or structuring email drafts for quick copy-and-pasting into other apps. Customizable outputs: Transform recordings into LinkedIn posts, podcast or video scripts, structured documents, or your own custom formats. Iterative refinement: Try different transformations of the same recording until you get exactly what you need. Multiple languages: Record in any of 90 languages, or record in one language and have the app translate your text into another. Offline and screen-off options: Record anywhere, even without Internet access. Try using background mode without your screen on. I often record with my AirPods while walking with my phone in my pocket. Founder’s tip: “Don’t confuse it with dictation,” says Letterly’s founder and CEO Anton Lebedev. “You don’t need to pronounce the perfect text you want to write. Instead, think out loud, speak slowly, quickly, or even chaotically. AI will understand you. Think of it like a writing assistant you’re telling what to write. The assistant can understand you and figure out how to rewrite the text.” Letterly Pricing: $80/year after a free trial Oasis Multi-purpose output: Get your recording transformed simultaneously into various formats—from a memo or outline to a blog post or TED talk. Make custom templates: Create and name short prompts that reflect your preferred styles or formats. Those become part of your personalized prompt library for transforming future recordings. I made one for my journal entries. Web accessibility: Like Letterly and Audiopen, you can access your recordings and transformed text through a browser on any device. Oasis pricing: $5/month or $50/year for enough credits for hundreds of monthly uses. AudioPen Customize rewrite length: Customize the length setting if you’d prefer summaries of your transcribed recordings to be shorter or longer. Create and access them on your phone or on any device through your browser. Shareable audio notes: Send individual audio note links to colleagues or collaborators. Or send then to other apps with a Zapier integration. Flexible organization: Combine multiple audio notes or their summaries into larger collections. You can search for old notes or arrange them in folders. Rich template selection: Choose from various transformation templates. AudioPen pricing: $99/year or $159/two years after a free trial. Bottom Line Start with Letterly if you want simplicity and reliability. Consider Oasis if you want a slightly cheaper option or need to simultaneously access multiple format variations of the same content. AudioPen is useful if you want to customize the length of your voice summaries or if sharing or combining audio notes is important to your workflow. Where to use voice AI Voice AI shines when typing isn’t practical or when you want to think freely without your hands on a keyboard. Here are situations where you can try it: At home Comfy chair: Capture book notes without interrupting your reading rhythm. Kitchen: Document recipe adjustments or cooking notes while your hands are busy with ingredients. Bedside: Record late-night musings without disrupting your wind-down routine with a bright screen. Garden: Log landscaping ideas or random thoughts while your hands are dirty. On the move Walking: Capture project ideas and inspiration during your daily stroll. Commute: Draft emails and plan your day while on the subway or bus. Car: Record thoughts safely after parking but before you forget an important idea. At work Quiet space: Create reflective journal entries while looking out the window. Conference: Capture insights between sessions to avoid being overwhelmed when you get home. Doctor’s office: Record appointment details and follow-up steps while the info is fresh. Active time Outdoors: Draft journal entries or creative ideas while surrounded by nature Exercise: Outline presentations or brainstorm on the treadmill Shopping: Create lists or remind yourself about products Voice AI on your laptop I used to rely exclusively on mobile voice AI apps, but lately I’ve been relying on laptop voice AI apps. These are less focused on transforming text and more on putting your spoken text on your clipboard so you can paste into any tool you’re using. It works with Google Docs, Word, email, or whatever else you’re using. I use these on my laptop because it’s quicker and easier for me to talk than to type. Here are three worth trying: Flow Quick to start: Once you’ve installed the software, just hold down the function key to start recording in any of 100+ languages. Your recording gets instantly transcribed and the cleaned-up text is copied to your clipboard. Works anywhere on your computer: Paste transcribed text directly into any application—email, documents, or messaging apps. Reduces screen and hand fatigue: Record while looking away from your screen to reduce eye strain and give your hands a break. Flow pricing: Free for up to 2,000 words/week; $12/month billed annually for unlimited words and extra features. $8/month for students and educators. TalkTastic Simple transcription: Made by the team that created the Oasis mobile app, TalkTastic is designed to be simpler. Instead of transforming your speech into various text types, it just puts a cleaned-up version of what you say onto your clipboard to paste into any app. Smart text transformation: You can optionally set it to analyze your screen context to offer transformed versions of your text. Free: While in beta, there’s no cost for TalkTastic. MacWhisper Advanced transcription: Use this free software to transcribe online meetings, podcasts, or live dictation. You can even upload files to transcribe. Pay once for pro features: Enable YouTube transcriptions, batch uploads, translation, and top AI model usage with a one-time purchase. MacWhisper pricing: Free for basic usage; about $60 for pro upgrade; 20% discount with this link. Journalists, students, or non-profits can email support@macwhisper.com for 50% off. Other ways to use your voice to benefit from AI ChatGPT has a powerful voice mode in its mobile and desktop apps. Rather than typing out AI queries, you can have a conversation with an AI bot. Here’s why that’s so useful. Perplexity’s mobile app voice AI mode is terrific. I ask it a series of questions, like an oracle. It beats Google on many of my queries. The AI understands what I’m asking, then gathers and summarizes a helpful response. Citations in the app ensure I can check on its info sources. Google’s Gemini and Microsoft’s Copilot have recently-upgraded mobile voice modes. Converse with human-sounding AI bots without thumb typing. Open-source options abound. This article is republished with permission from Wonder Tools, a newsletter that helps you discover the most useful sites and apps. Subscribe here. View the full article

A March 2025 survey by the National Association for Business Economics shows growing inflation concerns — and a diminishing chance of rate cuts — in 2025, but also optimism about avoiding a recession. View the full article

Elon Musk’s company, like the US itself, seems like a sell View the full article

US has grown to nearly two-thirds of global equity market value, but some analysts see danger in ‘huge bet on AI’View the full article

Final price expected to be £119mn, with work completed next year View the full article

Ukrainians want peace more than anyone else. But we demand real peaceView the full article

Staley is appealing against a lifetime FCA ban from City roles over his disclosure of links to the disgraced financier View the full article

Spending on defence will need to rise substantially if the UK is to meet the challenges it facesView the full article

Top regulator was concerned about German lender’s credit risk managementView the full article

Chevrolet Silverado’s complex supply chain leaves it particularly vulnerable to a looming trade warView the full article

Recommendations by parliamentarians come as ministers consider amendments to flagship legislation View the full article

In the business world, advertisers are the stunt performers. Our fragmented media and pop cultural landscape has forced brands to really push stunts into the weird and wonderful. Whether it’s Snoop threatening to give up smoking, Ben Affleck working a Dunkin’ drive-thru, or a devil baby terrorizing the streets of New York. Meanwhile in Hollywood, the stunt performers are the ones who actually pull off the death-defying action that can make us gasp. They’re a breath of IRL fresh air in a world blanketed by visual-effects technology. Now, for the biggest night in entertainment, these two worlds converge for a pretty epic stunt by both worlds’ definition. Disney Advertising, Jimmy Kimmel’s Kimmelot, and Ryan Reynolds’s Maximum Effort, have enlisted five different brands to create six commercials that will air during the Oscars ceremony, tonight at 4 p.m. PT/7 p.m. ET, featuring more than 75 stunt performers executing classic Hollywood stunts like skydiving, high falls, and dynamic fight scenes. John Campbell, Disney Advertising’s senior vice president of entertainment and streaming solutions, says that a lot of their conversations with CMOs have revolved around looking to create quality content that can maximize a given cultural moment, in particular to live audiences. The Oscars ticked all of those boxes. For the participating brands—Carnival Cruise Line, Kiehl’s, L’Oréal Paris, MNTN, and Samsung—Campbell says they saw the advantage of teaming up on a unique concept. “We had this concept to shine a light on the stunt community and the tremendous impact that they have on film for all of us as fans,” says Campbell. “We see them as Hollywood’s hidden heroes, so we put the hypothesis out there: What if we invited them to step into the spotlight on entertainment’s biggest night? And honestly, what came about is kind of wild. We had 75 real-life stunt performers, 150 crew members, and we did this in a little over a week of filming and producing.” The result may well be the biggest advertising stunt ever done for the Oscars. Brands playing nice Marketers typically do not like sharing the spotlight. But here, with Carnival Cruise Line, Kiehl’s, L’Oréal Paris, MNTN, and Samsung, there’s a balance among brands across product categories so that they can creatively all row in the same direction. Campbell says that the company has brands it works with on a consistent basis and had a sense of who might want to test the waters of a new concept like this. “The question is, are brands really going to play together? So we’ve had to find the right brands who are going to trust Disney advertisers, going to trust the Academy in order to say, ‘You know what, we are going to play together, and something really special is going to come about.‘” Kiehl’s general manager John Reed says this is the brand’s national TV-commercial debut, and it wanted to showcase authenticity, craftsmanship, and innovation. “With this being a multibrand project, it was important to us that the Kiehl’s spot felt endemic to the brand while fitting into the larger storytelling,” says Reed. “That we can stand out while fitting in.” L’Oreal Paris USA president Laura Branik says this idea was a natural fit to showcase the performance of the brand’s Infallible 3-Second Setting Mist in a breakthrough way. “The collaborative process was really rooted in a shared vision and creativity,” says Branik. “We worked closely with Disney Advertising from the very start—from brainstorming concepts to shaping the final execution. We have all been working toward the same goal of elevating stunt performers who are so deserving of this spotlight, which has made the process truly collaborative and fun.” The Carnival Cruises ad features stuntwoman Hannah Betts jumping from a helicopter at 11,000 feet right into the pool of a Carnival Cruise ship. The clip in the ad was Bett’s first take. “The Oscars delivers the right audience of prospective Carnival cruisers who love the fun of travel and live events,” says Carnival CMO Amy Martin Ziegenfuss. “This collaboration provided an opportunity to be more contextually relevant within the program, alongside other great brands and partners.” The only nonconsumer brand of the lineup is ad tech firm MNTN, the parent company for Reynolds’s creative shop Maximum Effort. CEO Mark Douglas says that the company aims to showcase the power of storytelling during commercial time, and this stunt aligns perfectly with that mission. “It’s the ultimate way to show how advertising can be as thrilling and impactful as the content it complements,” says Douglas. Stunt Advertising Evolved It’s a novel concept getting five brands to collaborate on a six-ad extravaganza during the Oscars to celebrate stunt performers. But this isn’t Disney Advertising’s first crack at experimental ads. Back in 2022, it worked with Kimmelot and Maximum Effort for the series finale of The Walking Dead. They created a series of ads for Autodesk, Deloitte, DoorDash, MNTN, and Ring, featuring four characters who died over the course of the show’s 11-season run. All five commercials were shot in two days. Last year, the companies collaborated again, this time for a Groundhog Day-like campaign for Lay’s. That was a series of eight ads starring Stephen Tobolowsky, who played Ned in the 1993 comedy-film classic Groundhog Day, just trying to buy a bag of chips. The spots ran 75 times on the calendar’s Groundhog Day, taking up a third of ABC’s commercial inventory for the day, appearing during Good Morning America, General Hospital, Shark Tank, 20/20, and Jimmy Kimmel Live!. Campbell says that the goal was to evolve the approach, this time with multiple brands, and on perhaps the biggest night of live-event TV outside of the Super Bowl. “It was really asking, how do we continue to push ourselves, to push the boundaries, and use the full platform of Disney in order to keep pushing these creative boundaries?” he says. The Stunts Chris Denison directed the Carnival spot and coordinated the stunts and cast every stunt performer across all six commercials. He’s performed stunts in films for Zac Efron, Jared Leto, and Ewan McGregor, as well as being Sam Worthington’s stunt double in James Cameron’s Avatar sequels. Denison says that his first thought when he heard about this overall idea was, “Dude, don’t mess this up!” “All kidding aside, when I first heard of the concept of a series of stunts-centric commercials geared toward paying homage to our profession, I was struck by a tremendous sense of responsibility, both to my peers and the legends of the business who came before us,” says Denison. “I knew instinctively that if the commercials that we produced were anything less than amazing, they would fall far short of the mission of honoring the overall stunt community. As a result, my team and I poured absolutely everything we had into this process.” The biggest challenge from a stunt perspective was using the action to tell a meaningful story inside the boundaries of a 30-second spot. “I’m a firm believer that action should be a storytelling device; that is, all stunts should be used to drive a story forward, else you risk losing your connection with the audience,” says Denison. “The creative team and our fantastic directors did us so many favors in this regard, but we as a department had to be absolutely ruthless about distilling the action down to its core.” He points to the Samsung “Fight” commercial. “Our first iteration of the fight was over a minute long,” says Denison. “With the help of our fearless fight coordinator, Steve Brown, we literally workshopped that thing for weeks, stripping out the proverbial shoe leather while highlighting the individual storytelling elements. I believe that the finished product is as compelling as a 30-second fight can get. This challenge persisted across each of the spots, and in every instance we put a hard focus on hyperefficient storytelling through cool action to create what we felt was the best result possible.” What stands out to Denison about this advertising stunt is the sheer variety in the spots. From a massive drop off a 12-story office building and a practical skydive out of a helicopter over an ocean to precision-driving in a trophy truck out in the desert. “Each spot is wildly different, and yet they all tell a broader story of what kind of content the stunt community is capable of producing,” says Denison. “It was incredibly fun getting to switch gears so rapidly and complete so many different action sequences in a short amount of time. As under the radar as stunt performers are used to flying, there are individuals within the stunt community whose names are even less well-known, and yet their work is absolutely everywhere on screen.” Denison goes on to point out a few of the less-than-famous folks Oscars viewers will see. “Whether or not the audience recognizes this, I think it’s a very fun Easter egg that the boom operator in the ‘Fight’ commercial is a UFC Hall-of-Famer, the first assistant director in the Kiehl’s spot is an accomplished supercross racer, and the dude in the floatie in the Carnival Cruise ship pool is Hugh Jackman’s stunt double.” View the full article

Journalists love a good tech prototype. These exciting experiments break up the monotony of covering the same incremental refreshes day-in and day-out.They also come with a downside: As attention grabbing as they are, there’s no guarantee that they’ll ever actually come to market. That’s why covering them can be a bit tricky, as it can be difficult to decide how much attention to give a product consumers might never actually see. Luckily, Lenovo actually made good on one of its earlier promises at CES this year, showing off a rollable laptop that’s set to launch sometime in 2025. Now, the company has three additional concepts it's showing off for Mobile World Congress, all of them arguably as cool. A foldable spin on that CES’ rollable screen Credit: Michelle Ehrhardt First, there’s the ThinkBook Flip AI PC, which you can think of as a more affordable take on that rolling laptop from CES. Rather than having some of the screen hidden inside the laptop and unfurling like a scroll, this laptop’s screen folds over the top-half of the device, stretching across both the sides of the lid. Using your hands, you can unfold it to get an especially tall vertical display, or, alternately, mirror the front of your screen to the backside for presentations. Or, you could use the laptop in a sort of tablet mode while it’s closed. It’s similar to existing devices like the Yoga Book 9i, sure, but unlike on those, the screen is continuous and the keyboard is built into the device. Credit: Michelle Ehrhardt There are upsides and downsides to this approach. Because there’s no need to house a rolling mechanism, the Flip AI PC is actually able to get a little taller than a rollable laptop, but at the same time, it’s also a bit more top heavy, so you can’t use the fully extended screen at a steep angle without extra support. Snap-on laptop monitors Credit: Michelle Ehrhardt Also in the realm of extended laptop displays are the Magic Bay Dual Display and 2nd Display laptop accessories. Technically, this isn’t the first time Lenovo’s shown off concepts for Magic Bay, existing technology that allows certain accessories to snap onto pogo pins on the back of some ThinkBook laptops. Those have even included secondary displays before, but this year’s ideas definitely are a bit more involved. Starting smaller, the Magic Bay 2nd Display is an 8-inch vertical screen that complements a laptop with a (non-touch) tablet-like companion. It’s similar to the 10-inch horizontal panel that sits above your laptop screen that Lenovo showed off at last year’s CES, and coincidentally recalls my husband’s own desktop setup. As handy as it seems—it’s even on a hinge for a customizable angle— it’s nonetheless overshadowed by Lenovo’s other Magic Bay display concept. Credit: Michelle Ehrhardt The Magic Bay Dual Display turns your laptop into a three-monitor workstation, with additional 13.3-inch screens on either side of your device, complete with a 120Hz refresh rate across both. At 2.6 pounds, it's rather heavy, so it comes with a built-in kickstand. This isn’t the first time I’ve seen an idea like this, but when I saw this in person, I was impressed by how much more intuitive the connection process seems compared to current third-party alternatives. A solar-powered laptop Credit: Michelle Ehrhardt Outside of the realm of screens, there’s also an experimental solar-powered laptop that can charge off sunlight, called the Yoga Solar PC. It works through a solar cell built into the back of the laptop’s lid, which Lenovo says has “an over 24% solar energy conversion rate” and can charge up enough to play one hour of video within about 20 minutes of direct sunlight. I didn’t get to see it in action, seeing as I was in a dark room while Lenovo was showing it off, but I can say the solar technology doesn’t seem like it’s impacting the form factor too much. It was plenty thin looking to my naked eye, and the company says it’s 0.6-inches thin and 2.69 pounds. If that’s still too big for you, Lenovo did also show off the Solar Power Kit for Yoga concept, which is a standard battery bank that connects to a detachable USB-C solar panel (I didn’t see this one in person). Credit: Lenovo Evolutions on earlier concepts Credit: Lenovo While I got to see (most of) the above concepts in person, the company also has a few iterative changes on earlier ideas, and showed that it’s prototyping out its own version of competing devices. These include a glasses-less 3D ThinkBook, similar to devices that Asus and Acer already have on the market, as well as an AI-powered robotic companion named Tiko that would attach to laptops via Magic Bay. For more functionality, there’s also a Tiko Pro in the works, a small horizontal Magic Bay display that combines the advantages of the 2nd Display with AI-powered widgets. Again, we’ve seen versions of these either from competitors or at previous Lenovo concept announcements, but they’re still worth keeping an eye on. A bit more exciting is the Hybrid Dimensional 24-Inch Curved Monitor, which can supposedly display 3D visuals while simultaneously showing 2D assets without any loss in resolution, although this needs to be seen with your own eye to really be judged, and I have yet to get hands-on time with it. Bringing AI to non-AI PCs Credit: Lenovo Finally, there are tools to bring AI to non-AI PCs. Lenovo says it’s experimenting with two dedicated neural processing units, or NPUs, that can connect to existing PCs either through a USB stick or a monitor. Working like an external graphics card, the Lenovo AI Stick contains a 32TOPS NPU that can plug into a computer without an NPU over Thunderbolt to allow it to run LLMs and AI graphics tools locally. It can work off just the laptop’s own power, but Lenovo says you can also plug it into a wall at the same time for increased performance. Meanwhile, the AI Display with NPU Inside does much the same thing, but rather than including the NPU in a hub-like structure, builds it right into your monitor. This comes with the added benefit of allowing the screen to physically rotate, elevate and tilt to follow your movements, but like most monitors, it’s much bulkier and does need to be plugged into a wall at all times. Lenovo also says it’s working on an “AI Ring” that would provide gesture-based control while in spatial computing apps, although the company didn’t go into much detail beyond that. And that’s it on Lenovo’s MWC concepts. It’s a pretty wide array. Again, none of these are guaranteed to come to market, but that does mean they get to be a bit wackier than the typical product announcement. Personally, I’d love to get my hands on the Magic Bay Dual Display, and I could see the AI Stick being useful to developers. In the meantime, the company also announced a number of chip upgrades for its existing ThinkPad, ThinkBook, IdeaPad, and Yoga laptops that will start hitting the market with the new Yoga 7 2-in-1 and IdeaPad Slim 3x this month. View the full article

Continent’s biggest powers gathered at London summit to salvage something from crumbling postwar orderView the full article

Europe will put troops on ground but plan ‘must have strong US backing’, says UK prime minister after London summitView the full article

OpenAI has introduced GPT-4.5, a research preview of its latest and most advanced AI model, which is now available to ChatGPT Pro users and developers worldwide. The model represents “a step forward in scaling up pre-training and post-training,” according to OpenAI. GPT-4.5 advances OpenAI’s approach to AI development by focusing on unsupervised learning, a technique that enhances a model’s ability to recognize patterns, draw connections, and generate creative insights without explicit reasoning. “GPT-4.5 is an example of scaling unsupervised learning by scaling up compute and data, along with architecture and optimization innovations,” OpenAI stated in its announcement. “The result is a model that has broader knowledge and a deeper understanding of the world, leading to reduced hallucinations and more reliability across a wide range of topics.” The model was trained on Microsoft Azure AI supercomputers, and OpenAI states that it is expected to “hallucinate less” compared to its predecessors. OpenAI notes that interacting with GPT-4.5 “feels more natural” and that the model has “improved ability to follow user intent, and greater ‘EQ.’” These advancements make it particularly useful for “tasks like improving writing, programming, and solving practical problems.” The company also highlights GPT-4.5’s ability to integrate ideas naturally in conversations. “GPT-4.5 has a better understanding of what humans mean and interprets subtle cues or implicit expectations with greater nuance and ‘EQ,’” the company stated. The model also “shows stronger aesthetic intuition and creativity.” According to OpenAI, GPT-4.5 has been evaluated against previous models, showing improved factual accuracy and reduced hallucination rates. In a comparative evaluation, OpenAI reported that GPT-4.5 scored a SimpleQA accuracy of 62.5%, higher than GPT-4o and other OpenAI models. Additionally, its hallucination rate was 37.1%, significantly lower than GPT-4o’s 61.8%. OpenAI also shared data indicating that human testers preferred GPT-4.5 over GPT-4o in several areas, including “creative intelligence,” “professional queries,” and “everyday queries.” As part of its release, OpenAI states that it conducted extensive safety evaluations on GPT-4.5. “Each increase in model capabilities is also an opportunity to make the models safer,” OpenAI said. The model was trained using “new techniques for supervision that are combined with traditional supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF).” The company also referenced its Preparedness Framework, which it followed to “stress-test our improvements” before deployment. GPT-4.5 is now available to ChatGPT Pro users on web, mobile, and desktop platforms. OpenAI announced that it “will begin rolling out to Plus and Team users next week, then to Enterprise and Edu users the following week.” The model supports various capabilities, including “search, file and image uploads, and the ability to use canvas to work on writing and code.” However, OpenAI clarified that GPT-4.5 “does not currently support multimodal features like Voice Mode, video, and screensharing in ChatGPT.” For developers, GPT-4.5 is also being made available through OpenAI’s API offerings, including the Chat Completions API, Assistants API, and Batch API. “Based on early testing, developers may find GPT-4.5 particularly useful for applications that benefit from its higher emotional intelligence and creativity—such as writing help, communication, learning, coaching, and brainstorming,” OpenAI stated. However, due to its high computational requirements, OpenAI is evaluating “whether to continue serving it in the API long-term.” The company encourages user feedback to help guide this decision. OpenAI acknowledges that GPT-4.5 is still in a research preview phase. “We’re still exploring what it’s capable of and are eager to see how people use it in ways we might not have expected,” the company stated. “With every new order of magnitude of compute comes novel capabilities,” OpenAI added. “We continue to be surprised by the creativity of the community in uncovering new abilities and unexpected use cases. With GPT-4.5, we invite you to explore the frontier of unsupervised learning and uncover novel capabilities with us.” This article, "OpenAI Unveils GPT-4.5, Its Most Advanced AI Model to Date" was first published on Small Business Trends View the full article

OpenAI has introduced GPT-4.5, a research preview of its latest and most advanced AI model, which is now available to ChatGPT Pro users and developers worldwide. The model represents “a step forward in scaling up pre-training and post-training,” according to OpenAI. GPT-4.5 advances OpenAI’s approach to AI development by focusing on unsupervised learning, a technique that enhances a model’s ability to recognize patterns, draw connections, and generate creative insights without explicit reasoning. “GPT-4.5 is an example of scaling unsupervised learning by scaling up compute and data, along with architecture and optimization innovations,” OpenAI stated in its announcement. “The result is a model that has broader knowledge and a deeper understanding of the world, leading to reduced hallucinations and more reliability across a wide range of topics.” The model was trained on Microsoft Azure AI supercomputers, and OpenAI states that it is expected to “hallucinate less” compared to its predecessors. OpenAI notes that interacting with GPT-4.5 “feels more natural” and that the model has “improved ability to follow user intent, and greater ‘EQ.’” These advancements make it particularly useful for “tasks like improving writing, programming, and solving practical problems.” The company also highlights GPT-4.5’s ability to integrate ideas naturally in conversations. “GPT-4.5 has a better understanding of what humans mean and interprets subtle cues or implicit expectations with greater nuance and ‘EQ,’” the company stated. The model also “shows stronger aesthetic intuition and creativity.” According to OpenAI, GPT-4.5 has been evaluated against previous models, showing improved factual accuracy and reduced hallucination rates. In a comparative evaluation, OpenAI reported that GPT-4.5 scored a SimpleQA accuracy of 62.5%, higher than GPT-4o and other OpenAI models. Additionally, its hallucination rate was 37.1%, significantly lower than GPT-4o’s 61.8%. OpenAI also shared data indicating that human testers preferred GPT-4.5 over GPT-4o in several areas, including “creative intelligence,” “professional queries,” and “everyday queries.” As part of its release, OpenAI states that it conducted extensive safety evaluations on GPT-4.5. “Each increase in model capabilities is also an opportunity to make the models safer,” OpenAI said. The model was trained using “new techniques for supervision that are combined with traditional supervised fine-tuning (SFT) and reinforcement learning from human feedback (RLHF).” The company also referenced its Preparedness Framework, which it followed to “stress-test our improvements” before deployment. GPT-4.5 is now available to ChatGPT Pro users on web, mobile, and desktop platforms. OpenAI announced that it “will begin rolling out to Plus and Team users next week, then to Enterprise and Edu users the following week.” The model supports various capabilities, including “search, file and image uploads, and the ability to use canvas to work on writing and code.” However, OpenAI clarified that GPT-4.5 “does not currently support multimodal features like Voice Mode, video, and screensharing in ChatGPT.” For developers, GPT-4.5 is also being made available through OpenAI’s API offerings, including the Chat Completions API, Assistants API, and Batch API. “Based on early testing, developers may find GPT-4.5 particularly useful for applications that benefit from its higher emotional intelligence and creativity—such as writing help, communication, learning, coaching, and brainstorming,” OpenAI stated. However, due to its high computational requirements, OpenAI is evaluating “whether to continue serving it in the API long-term.” The company encourages user feedback to help guide this decision. OpenAI acknowledges that GPT-4.5 is still in a research preview phase. “We’re still exploring what it’s capable of and are eager to see how people use it in ways we might not have expected,” the company stated. “With every new order of magnitude of compute comes novel capabilities,” OpenAI added. “We continue to be surprised by the creativity of the community in uncovering new abilities and unexpected use cases. With GPT-4.5, we invite you to explore the frontier of unsupervised learning and uncover novel capabilities with us.” This article, "OpenAI Unveils GPT-4.5, Its Most Advanced AI Model to Date" was first published on Small Business Trends View the full article

Over the past week, we have seen two very different sides to the president’s communication styleView the full article

President’s comments about Solana, XRP and Cardano boost market after recent shake-out View the full article

Remote work offers flexibility and freedom, but it comes with a unique set of challenges. How do you build trust when you can’t see your team? How do you provide accountability without creating an atmosphere of surveillance? How do you keep everyone in sync without communication constantly disrupting the workday? The lack of in-person interactions The post Building trust in remote teams: The role of transparent productivity tools appeared first on RescueTime Blog. View the full article

President Masoud Pezeshkian on defensive after impeachment of Abdolnaser HemmatiView the full article

Despite a traumatic beginning of the year from the devastating Eaton and Palisades fires, Los Angeles is ready to celebrate one of its largest exports: movies. Hollywood’s biggest bash, the 97th Academy Awards presented by the Academy of Motion Picture Arts and Sciences, will air tonight, March 2, at 4 p.m. PT / 7 p.m. ET. The action is taking place at the Dolby Theatre at Ovation Hollywood. Let’s get up to speed so you can watch like an A-lister: How the L.A. fires impacted the 2025 Oscars The fires delayed the Oscar nominations announcement twice and extended the nominations voting period. And though the actual awards ceremony date was not affected, the annual Oscars Nominees Luncheon was canceled out of sensitivity to the Southern California community, and the Scientific and Technical Awards ceremony had to be rescheduled from February 18 to April 29. Notable 2025 Oscar nominees Emilia Pérez dominates the Oscars with 13 nominations, making it the most nominated non-English movie in Oscar history. Jacques Audiard’s crime musical was once considered a tight contender for best picture, but controversy surrounding its star Karla Sofía Gascón’s raft of offensive tweets is expected to impact its chances (although it just won best film and best director at the Cesar Awards, France’s equivalent to the Oscars, on Friday in Paris). Gascón is the first openly trans actor to be nominated for an Oscar. Tied for second place with 10 nominations is Brady Corbet’s immigrant story The Brutalist and Jon M. Chu’s musical Wicked, both nominated for best picture. This is the first time two musicals have been in the best picture category since 1969 with Oliver! and Funny Girl. (Oliver! won.) With eight nominations each, A Complete Unknown and Conclave are also up for best picture. You can find a complete list of the 2025 Oscar nominees on the Academy’s website. Who is hosting the 2025 Oscars? Comedian, former late-night host, and podcaster Conan O’Brien is taking on hosting duties this year for the first time. The two-time Emmy Awards host revealed that he’s never even attended the Academy Awards before. “I only agreed to host so that I could get invited,” O’Brien said (jokingly?) during a news conference Wednesday. He will get backup from presenters including Oprah Winfrey, Ben Stiller, Sterling K. Brown, Willem Dafoe, John Lithgow, and Amy Poehler. Keeping with tradition, last year’s top acting winners, Robert Downey Jr., Cillian Murphy, Da’Vine Joy Randolph, and Emma Stone will also present. Who are the musical performers? Historically, the best original song nominees were performed during the ceremony. This year, the Academy is shaking things up and having songwriters share their personal reflections instead. That doesn’t mean the music stops. Wicked’s Cynthia Erivo and Ariana Grande will open up the night with a musical performance. Lisa of Blackpink, Queen Latifah, Doja Cat, and Raye are also cooking up something special. Who are some favorites to win this year? There have been lots of twists and turns leading up to the big night. Some categories, such as best supporting actor and actress, feel almost inevitable. Kieran Culkin and Zoe Saldaña will almost certainly walk away with Oscars in these categories. Best picture, actor, and actress are more unpredictable. In the actor race, Oscar-winner Adrien Brody seemed the front-runner until last weekend’s SAG Awards upset when Timothée Chalamet took home the honor for his portrayal of Bob Dylan. Traditionally, the Oscars don’t like to honor younger artists, but the times could be a-changin’. In the actress category, Demi Moore of The Substance and Mikey Madison of Anora have split the precursor awards. Moore’s heartfelt speeches at SAG, Critics Choice, and the Golden Globes ceremonies charmed audiences. But don’t count Madison out; she took home the BAFTA and Independent Spirit Awards trophies for best actress. Further complicating the issue, some critics are predicting a surprise third choice, I’m Still Here’s Fernanda Torres. This is truly anyone’s race. Best picture is also uncertain. Sean Baker’s Anora seemed to be a safe bet because of its PGA, DGA, and WGA wins, but Conclave’s SAG and BAFTA wins might make a case for Edward Berger’s papal thriller. How can I watch or stream the 2025 Oscars? There are many ways to see what film comes out on top. The most straightforward for those with traditional cable TV subscriptions is to tune into ABC at 4 p.m. PT / 7 p.m ET. You can also watch it for free on ABC with an over-the-air antenna. For the first time ever, cord-cutters can also stream the Oscars live directly on Hulu, which is offering a free trial for new subscribers. Or catch the show on one of the live-TV streaming services that offer ABC as part of a bundle: Sling TV (in some markets) Fubo TV YouTube TV Hulu + Live TV View the full article

High interest rates and uncertainty over immigration policy discourage homeowners from selling or renovating propertiesView the full article