Get ready for slick light strips and futuristic lightcycles.Disney just released the first trailer for Tron: Ares, the long-planned Tron: Legacy sequel. The minute-and-a-half trailer doesnt say much about the story but shows plenty of the movies visuals, which look dark, moody, and filled with the series signature light trails.The trailer opens in the physical world at night, as Jared Letos Ares, a Program made physical, flees from police on a light cycle, slicing one in half using his light trail as a weapon. The shots that follow show a massive airship hovering over the real-world city, visible only by the red light strips on its outside. The rest has people looking on in horror at the airship, dogfights between human aircraft and fighters from the Tron digital world, and what looks like a clip of Ares being given his physical body. All of that is set to the music of Nine Inch Nails, which is handling the soundtrack this time around. It ends with a voiceover from Jeff Bridges, reprising his role as Kevin Flynn and saying, Ready? Theres no going back. The movie hits theaters on October 10th.Disney included the poster above in an email to The Verge announcing the trailers release. In a YouTube video from Thursdays CinemaCon presentation about Ares, Leto said his character is a highly advanced program who has entered the real world on a do-or-die mission to fulfill his directive, and promised that the movie will hit you right in the grid wherever that is. In addition to Leto and Bridges, Tron: Ares is directed by Joachim Rnning and its stars include Gillian Anderson, Greta Lee, Evan Peters, Hasan Minhaj, Jodie Turner-Smith, Arturo Castro, and Cameron Monaghan.

Were nearing the end of the 2025 Mens March Madness Tournament. In perhaps one of the least interesting outcomes of all time, all four top seeds made it to the semi-finals. So, if you happened to base your entire bracket on the number 1, congratulations! With the rest of the tournament playing out over the next couple of days, it might not be worth picking up a new subscription or cable plan. But that shouldnt mean you have to miss out on the fun. With the magic of free trials, heres how you can watch the March Madness Final Four games online for free. How to Watch the Final Four Games for FreeParamount+See it at Paramount+The rest of the Mens March Madness tournament takes place today and Monday. That means you only need three days of a streaming service to watch the semi-finals and the National Championship. All three games will be airing live on CBS and simultaneously streamed on Paramount+, which happens to offer a 7-day free trial. In other words, if you sign up for Paramount+ right now, you can watch the Final Four games as well as the National Championship before your trial ends. Alternatively, you could use an HDTV antenna to watch the games live via your local CBS channel.Other Free Trials That Let You Watch March MadnessParamount may have the longest free trial (and cheapest ongoing subscription cost if you forget to cancel it), but plenty more live TV subscriptions include CBS. Services like Hulu + Live TV and DirecTV Stream might make sense for anyone who wants ongoing sports coverage, including NBA livestreams.3-Day Free TrialHulu + Live TVSee it at Hulu5-Day Free TrialDirecTV StreamSee it at DirecTV7-Day Free TrialFubo TVSee it at Fubo21-Day Free TrialYouTube TVSee it at YouTubeMarch Madness Final Four Game ScheduleThere's only three games left in the March Madness tournament, all of which are being played in San Antonio at the Alamodome. See the full tournament schedule at the NCAA website. Final Four (Semi-Finals) - Saturday, April 5(1) Florida vs. (1) Auburn - 6:09pm (CBS, Paramount+)(1) Duke vs. (1) Houston - 8:49pm (CBS, Paramount+)National Championship - Monday, April 7 TBA - 8:50pm (CBS, Paramount+) All Top Seeds, But a Range of Experience in the Final FourEvery one of these top-seeded teams has something to prove. This is Duke's first time reaching the Final Four since Coach K retired back in 2022, and the Blue Devils lead predictions for taking the crown. Florida won back-to-back championships in 2006 and 2007, though the last time they made it to the Final Four was in 2014. Meanwhile, this is only Auburn's second time ever making it to this stage. Houston actually holds the record for the most Final Four appearances without any Championship wins (seven total), most recently reaching the semi-finals in 2021. Blythe (she/her) is an SEO Coordinator at IGN who spends way too much time in character customization screens and tracking down collectibles.

In the Apple Silicon era, MacBooks have become more affordable than ever with brand new models starting as low as $649. With looming tariffs, these great deals could potentially be coming to an end very soon so if youre in the market for a new MacBook for any reason, now may be one of the best ever times to purchase.M1 MacBook AirApple has launched 4 MacBook Airs since entering the Apple Silicon era, with the M1, M2, M3, and M4 chips. The M1 MacBook Air still utilizes a more familiar wedge design from the Intel era, although a lot of users might prefer that.The M1 MacBook Air, despite being four years old, remains an excellent laptop for its price. It boasts blazing fast performance thanks to the M1 chip, a stunning retina display (though with thicker bezels compared to the newer MacBook Airs), a great keyboard, up to 18 hours of battery life, and it doesnt sound like an airplane thats about to take off when you open it in public.Since the M1 MacBook Air is the oldest on this list, its also the cheapest. One caveat with the M1 MacBook Air is that its harder to find new ones with more than 256GB of storage, so if you need a lot of storage, youll have to buy a newer Air. On the bright side, it does remain to be the only Apple Silicon MacBook to come in a proper Gold color.You can buy the M1 MacBook Air for as low as $649 at Walmart in all 3 colors: Silver, Space Gray, and Gold.13-inch MacBook Air (M1) with 256GB of storage, 8GB of memory: $649 at WalmartM3 MacBook AirIn 2022, Apple introduced the redesigned MacBook Air with a thinner design, a bigger display with slimmer bezels, and reintroduced MagSafe. It also came in new Starlight and Midnight colors in addition to Silver and Space Gray, although Gold was no longer. This fresh new design was introduced with the M2 chip, and its lived on since. Unfortunately, M2 chip models have effectively been cleared out at every major retailer, so the deals will start with M3 chip models. Apple introduced an M4 chip model earlier this year, but as long as cutting-edge performance isnt your priority, this model will suit you perfectly fine.Right now, you can secure up to $300 off, and there are configurations available with 512GB of storage. There are also deals available on the 15-inch version of the M3 MacBook Air.13-inch models with M3 chip:13-inch MacBook Air with 256GB of storage, 16GB of memory: $899 at Amazon ($200 off!)13-inch MacBook Air with 512GB of storage, 16GB of memory: $999 at Amazon ($300 off!)13-inch MacBook Air with 512GB of storage, 24GB of memory: $1199 at Amazon ($300 off!)15-inch models with M3 chip (limited availability):15-inch MacBook Air with 512GB of storage, 24GB of memory: $1399 at Amazon ($300 off!)Some deals require clipping an on-page coupon. Nonetheless, some of the best deals come in on higher trims, allowing you to get higher-end specs at a much more affordable price.M4 MacBook AirLast month, Apple announced a refreshed MacBook Air with M4 chip. This computer is mostly similar to the previous model, except it now features a new Center Stage webcam and a new Sky Blue color that replaces Space Gray.This computer also came in with a lower starting price, allowing for there to be some deal-like prices despite the computer being the latest and greatest.13-inch models with M4 chip:13-inch MacBook Air with 256GB of storage, 16GB of memory: $949 at Amazon13-inch MacBook Air with 512GB of storage, 16GB of memory: $1069 at Amazon13-inch MacBook Air with 512GB of storage, 24GB of memory: $1381 at Amazon15-inch models with M4 chip:15-inch MacBook Air with 256GB of storage, 16GB of memory: $1139 at Amazon15-inch MacBook Air with 512GB of storage, 16GB of memory: $1364 at Amazon15-inch MacBook Air with 512GB of storage, 24GB of memory: $1549 at AmazonWhile these prices are all under $100 off of the typical retail price, its still worth highlighting. With tariffs, price hikes on new Apple products could happen in the near future.14-inch MacBook ProApple redesigned the MacBook Pro in 2021 with a brand new 14-inch miniLED display with ProMotion, more ports (most notably an HDMI port and SD card reader), as well as bigger batteries and better thermals. It also gained a 1080p webcam for the first time.Although the 14-inch MacBook Pro isnt as affordable as the other MacBooks listed, if you want that amazing display and the extra ports, there are still some good deals to find. All of the models listed below are M4 and M4 Pro chip models, providing the latest and greatest performance and features.M4 chip models:M4 MacBook Pro with 512GB of storage, 16GB of memory: $1429 at Amazon ($170 off!)M4 MacBook Pro with 1TB of storage, 16GB of memory: $1619 on Amazon ($180 off!)M4 MacBook Pro with 1TB of storage, 24GB of memory: $1799 on Amazon ($200 off!)M4 Pro MacBook Pro with 512GB of storage, 24GB of memory: $1789 on Amazon ($210 off!)My favorite Apple accessories on Amazon:Follow Michael:X/Twitter,Bluesky,InstagramAdd 9to5Mac to your Google News feed. FTC: We use income earning auto affiliate links. More.Youre 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. Dont know where to start? Check out our exclusive stories, reviews, how-tos, and subscribe to our YouTube channel

Two senators have sent a letter to multiple AI companion companies requesting information about their safety practices, including details about internal safety assessments and timelines of the implementation of guardrails, asCNNreported yesterday.The action follows the filing of two high-profile child welfare lawsuits against the Google-tied chatbot startup Character.AI, which has been accused in court filings by three families of facilitating the sexual and emotional abuse of minor users, allegedly resulting in severe mental and emotional suffering, violence, behavioral changes, and one death. (Google and Character.AI cofounders Noam Shazeer and Daniel de Freitas are also named as defendants in the lawsuits.)Drafted by democratic senators Alex Padilla of California and Peter Welch of Massachusetts, the letter cites details of the accusations against Character.AI as cause for alarm, calling specific attention to concern for minor users of AI companion apps and directly referencing Sewell Setzer III, a 14-year-old user of Character.AI who died by suicide in February 2024 after extensive and extraordinarily intimate interactions with the platform's anthropomorphic chatbots."We write to express our concerns regarding the mental health and safety risks posed to young users of character- and persona-based AI chatbot and companion apps," reads the letter, "including Character.AI."Setzer's death, first reported by The New York Times, made headlines in October after his mother, Megan Garcia, filed the first of the two ongoing child welfare suits against the chatbot startup. The second complaint was filed in Texas in December on behalf of two more families whose minor kids are said to have experienced significant mental and physical harm as a result of using the service. One of them, who was 15 when he started using the app, began physically self-harming after discussing self-injury with a Character.AI bot.Both lawsuits, which together argue that Character.AI and its benefactor Google knowingly released a dangerous and untested product into the marketplace, have made waves amongst the public, particularly parents. And now, it looks like lawmakers on Capitol Hill are paying attention."In light of recent reports of self-harm associated with this emerging application category, including the tragic suicide of a 14-year-old boy," the letter continues, asking that recipients "respond in writing outlining what steps you are taking to ensure that the interactions taking place on your products between minors and your artificial intelligence tools are not compromising the mental health and safety of minors and their loved ones."Per a press release, the letter was sent to Character.AI, Chai Research Corp, and Replike maker Luka, Inc.Replika, which has been a player in the digital companion space for many years, is currently facing a Federal Trade Commission complaint from advocacy groups alleging that it's engaged in deceptive marketing practices aimed at hooking vulnerable users. Other Replika controversies include its alleged role in encouraging a troubled young man in the UK to attempt to assassinate the late Queen Elizabeth II with a crossbow, as well as men using the app to abuse virtual girlfriends.All three companies offer a version of a similar product: access to emotive, lifelike chatbots designed to embody specific personas. (Think characters like "goth girlfriend," ersatz AI versions of celebrities, fake therapists or other professionally-styled bots, or virtually any fictional character in existence.)In some cases, users carry on extensive fictional roleplays with the bots; others treat the characters like trusted confidantes, with users frequently developing emotional, romantic, or sexual relationships with the characters.But while companion apps have proliferated amid the AI boom, experts have consistently warned that the same design features that make them so engaging their penchant for sycophancy and flattery, always-on availability, and human-like tenor, to name a few may put vulnerable users at a heightened risk for harm.The senators' letter speaks to these concerns, writing that the "synthetic attention" such bots give to users "can, and has already, led to dangerous levels of attachment and unearned trust stemming from perceived social intimacy." This trust, they add, can "cause users to disclose sensitive information about their mood, interpersonal relationships, or mental health, which may involve self-harm and suicidal ideation complex themes that the AI chatbots on your products are wholly unqualified to discuss."The lawmakers are requesting a few different pieces of information from the AI companies. They first ask that the firms provide them with the "current and historical" safety guardrails enacted in their products and, importantly, a timeline of their implementation. (Character.AI, for example, has historically been extremely reactive to apparent gaps in safety guardrails, repeatedly promising to add new safety features after controversies arise.)The senators are also requesting that the companies provide them with information about the data used to train their AI models, and how that training material "influences the likelihood of users encountering age-inappropriate or other sensitive themes."The companies are further asked to disclose details about safety personnel, as well as a description of the services and support provided to safety-oriented staffers like content moderators and AI red-teamers, whose work often necessitates contact with sensitive or disturbing material.Luka and Chai did not respond to CNN's request for comment. In a statement, Character.AI told CNNthat the company takes the senators' concerns "very seriously.""We welcome working with regulators and lawmakers," the company added, "and are in contact with the offices of Senators Padilla and Welch."As it stands, like other generative AI firms, these companion companies have operated within a largely unregulated federal landscape. To that end, the senators' letter is still exploratory, and is among the first steps by Capitol Hill lawmakers to investigate the safety measures and perhaps more tellingly, founding safety practices and principles at industry leaders like Character.AI, Replika, and Chai.But it's a step nonetheless. And given that Character.AI, in particular, has repeatedly declined to provide us with information about how it's assessed the safety of its platform for minor users, we'll be paying close attention to what happens next.Share This Article

Find hints and answers for the New York Times' Connections puzzle for Sunday, April 6.

CLOSE X By Alyce Collins Life and Trends Reporter A recruiter has shared a stark warning about the risks of AI after discovering she wasn't interviewing a real personbut a fake AI candidate.Several weeks ago, Bettina Liporazzi, the recruiting lead at letsmake.com was contacted by a seemingly ordinary candidate who was looking for a job. Their initial message was clearly AI-generated, but Liporazzi told Newsweek that this "didn't immediately raise any flags" because that's increasingly commonplace.However, what did pique her interest was that another candidate reached out approximately an hour prior with an almost identical message. This gave her some initial doubts, but she arranged an interview with the candidate to try and get a better idea about them."They provided a name and an email to schedule the call, but when they joined, they weren't logged in to any email address," Liporazzi, of Buenos Aires, Argentina, said. "They claimed their camera was broken, which is a common excuse. I insisted that having the camera on was required, and they left and rejoined the call twice, saying they had to restart their laptop."Each time the candidate joined the call, Liporazzi got a warning from Google to say the person wasn't signed in and "might not be who they claim to be."The AI candidate that Bettina Liporazzi found herself interviewing before they abruptly left the call.The AI candidate that Bettina Liporazzi found herself interviewing before they abruptly left the call.Bettina Liporazzi / letsmake.com On the third attempt, the candidate finally put their camera on, and Liporazzi's suspicions were confirmed."I noticed they were using an AI filter to hide their real identity. So, I asked them to place a hand in front of their face as it would remove their AI cover. They didn't do it and left the call a few seconds later," she continued.Overall, it was a "creepy experience" for Liporazzi, but she shared footage of the video call on LinkedIn to educate others about the risks of AI. The video shows the moment Liporazzi asked the AI candidate to put a hand in front of their face, only for them to leave abruptly.The encounter left many people stunned, and the post has amassed over 3,600 reactions of 800 comments since it was shared online.As generative-AI becomes increasingly powerful, the line between what's real and fake is becoming harder to decipher. Ben Colman, co-founder and CEO of Reality Defender, a deepfake detection company, tells Newsweek that AI impersonation in recruiting is "just the tip of the iceberg."Colman explained that recruiters need "practical strategies to verify authentic human interaction" during calls. However, relying solely on humans to solve this won't work, as Colman notes that even his own staff can sometimes struggle to do this manually."Look for inconsistencies in facial movements, particularly around the eyes and mouth, unnatural voice cadence, or delayed responses to unexpected questions. The most effective approach combines technical verificationlike requesting candidates to perform specific real-time actionswith organizational protocols that incorporate detection tools," Colman said.Having a "layered security approach" is critical, according to Colman, as he says that deepfake technology is becoming even less perceptible.This was echoed by Liporazzi, who is concerned by the number of fake online accounts, and she urges others to stay vigilant when hiring. She felt well-equipped to handle the situation because letsmake.com uses AI to aid efficiency, meaning she was aware of the dangers of identity fraud.Liporazzi's alternative ways of spotting an "AI imposter" include asking them to turn their face to the side, ensuring their camera is on during the call, and asking them to remove filters and backgrounds."I was surprised by how many people had no idea this was happening, and I'm glad that sharing my experience has helped others who might have fallen for identity fraud. If it raises awareness and helps others stay alert, then at least something good has come from it," Liporazzi told Newsweek."There are resources online that help recruiters recognize what a fake account looks like. I encourage recruiters to do everything in their power to stay ahead of it by subscribing to newsletters, networking with professionals, and staying up to date with emerging technologies and verification tools," she continued.Among the hundreds of comments on her LinkedIn post, many people shared their shock over the AI interview candidate.One comment reads: "Wow, I did not know this was even a thing!"Another LinkedIn user wrote: "Props to you Bettina Liporazzi for spotting this quickly and knowing what to do! Appreciate you sharing."One person replied: "Take me back to a pre gen AI world please."Have you had a workplace dilemma? Let us know via life@newsweek.com. We can ask experts for advice, and your story could be featured on Newsweek. Is This Article Trustworthy?Is This Article Trustworthy?Newsweek is committed to journalism that is factual and fairWe value your input and encourage you to rate this article.Newsweek is committed to journalism that is factual and fairWe value your input and encourage you to rate this article.Slide Circle to VoteReader Avg.NoModeratelyYes VOTE

Meta has released a new collection of AI models, Llama 4, in its Llama family on a Saturday, no less. There are four new models in total: Llama 4 Scout, Llama 4 Maverick, and Llama 4 Behemoth. All were trained on large amounts of unlabeled text, image, and video data to give them broad visual understanding, Meta says. The success of open models from Chinese AI lab DeepSeek, which perform on par or better than Metas previous flagship Llama models, reportedly kicked Llama development into overdrive. Meta is said to have scrambled war rooms to decipher how DeepSeek lowered the cost of running and deploying models like R1 and V3.Scout and Maverick are openly available on Llama.com and from Metas partners, including the AI dev platform Hugging Face, while Behemoth is still in training. Meta says that Meta AI, its AI-powered assistant across apps including WhatsApp, Messenger, and Instagram, has been updated to use Llama 4 in 40 countries. Multimodal features are limited to the U.S. in English for now.Some developers may take issue with the Llama 4 license. Users and companies domiciled or with a principal place of business in the EU are prohibited from using or distributing the models, likely the result of governance requirements imposed by the regions AI and data privacy laws. (In the past, Meta has decried these laws as overly burdensome.) In addition, as with previous Llama releases, companies with more than 700 million monthly active users must request a special license from Meta, which Meta can grant or deny at its sole discretion.These Llama 4 models mark the beginning of a new era for the Llama ecosystem, Meta wrote in a blog post. This is just the beginning for the Llama 4 collection. Image Credits:MetaMeta says that Llama 4 is its first cohort of models to use a mixture of experts (MoE) architecture, which is more computationally efficient for training and answering queries. MoE architectures basically break down data processing tasks into subtasks and then delegate them to smaller, specialized expert models.Maverick, for example, has 400 billion total parameters, but only 17 billion active parameters across 128 experts. (Parameters roughly correspond to a models problem-solving skills.) Scout has 17 billion active parameters, 16 experts, and 109 billion total parameters.According to Metas internal testing, Maverick, which the company says is best for general assistant and chat use cases like creative writing, exceeds models such as OpenAIs GPT-4o and Googles Gemini 2.0 on certain coding, reasoning, multilingual, long-context, and image benchmarks. However, Maverick doesnt quite measure up to more capable recent models like Googles Gemini 2.5 Pro, Anthropics Claude 3.7 Sonnet, and OpenAIs GPT-4.5.Scouts strengths lie in tasks like document summarization and reasoning over large codebases. Uniquely, it has a very large context window: 10 million tokens. (Tokens represent bits of raw text e.g. the word fantastic split into fan, tas and tic.) In plain English, Scout can take in images and up to millions of words, allowing it to process and work with extremely lengthy documents. Scout can run on a single Nvidia H100 GPU, while Maverick requires an Nvidia H100 DGX system or equivalent, according to Metas calculations.Metas unreleased Behemoth will need even beefier hardware. According to the company, Behemoth has 288 billion active parameters, 16 experts, and nearly two trillion total parameters. Metas internal benchmarking has Behemoth outperforming GPT-4.5, Claude 3.7 Sonnet, and Gemini 2.0 Pro (but not 2.5 Pro) on several evaluations measuring STEM skills like math problem solving. Of note, none of the Llama 4 models is a proper reasoning model along the lines of OpenAIs o1 and o3-mini. Reasoning models fact-check their answers and generally respond to questions more reliably, but as a consequence take longer than traditional, non-reasoning models to deliver answers.Image Credits:MetaInterestingly, Meta says that it tuned all of its Llama 4 models to refuse to answer contentious questions less often. According to the company, Llama 4 responds to debated political and social topics that the previous crop of Llama models wouldnt. In addition, the company says, Llama 4 is dramatically more balanced with which prompts it flat-out wont entertain.[Y]ou can count on [Lllama 4] to provide helpful, factual responses without judgment, a Meta spokesperson told TechCrunch. [W]ere continuing to make Llama more responsive so that it answers more questions, can respond to a variety of different viewpoints [] and doesnt favor some views over others.Those tweaks come as some White House allies accuse AI chatbots of being too politically woke. Many of President Donald Trumps close confidants, including billionaire Elon Musk and crypto and AI czar David Sacks, have alleged that popular AI chatbotscensor conservative views. Sacks has historicallysingled outOpenAIs ChatGPT as programmed to be woke and untruthful about political subject matter.In actuality, bias in AI is an intractable technical problem. Musks own AI company, xAI, hasstruggledto create a chatbot that doesnt endorse some political views over others.That hasnt stopped companies including OpenAI from adjusting their AI models to answer more questions than they would have previously, in particular questions relating to controversial subjects.

When the Xiaomi 15 Ultra launched with its new periscope lens and stacked camera setup earlier this year, I just knew I had to get the device in hand.

Masayoshi Soken on his life in video games from the soundtrack that inspired him to the titles hes currently playing.

After more than a decade in development, Middle-earth Extended Edition has launched into version 1.0 for The Lord of the Rings: Battle for Middle-earth. Unfortunately, Battle for Middle-earth isnt available to purchase digitally, although you can find it used on eBay for the price of a new PlayStation 5 title. This mod has been a long time coming, but it adds the finishing touches on a project that has expanded a beloved RTS far beyond its original scope.Battle for Middle-earth was developed by Danger Close Games and released in December 2004. Middle-earth Extended Edition started development eight years later in 2012, and its developer Rohirim91 released the mod piecemeal as they completed different aspects of it. Todays release is the first in four years, but it more or less marks the mod as complete. That said, Rohirim91 hasnt stated whether they intend to continue to work on the mod. Recommended VideosThe release isnt marked with celebratory language, but instead reads as a routine update. Rohirim91 provides instructions on how to get the mod to run and writes, Middle-earth Extended Edition returns with an expanded set of playable maps, new features, polished campaigns and AI, balance changes and bugfixes. BFME 1 Patch 1.06 required.Image used with permission by copyright holderThe mod is a prime example of what makes gaming such an engrossing hobby. Its a work of passion, forged over the span of 13 years for a game thats now 21 years old. While you cant buy Battle for Middle-earth digitally, there has been a petition to bring it to GOG. If youre interesting in lending your voice to that effort, shoot the GOG team an email and let them know youd like to see it.RelatedAnd if youre a fan of this old-school RTS, you can download the most recent version of the mod here. It adds multiple factions, new maps, and tweaks the enemy AI to be much improved over its original version. In many ways, it breathes new life into a game that can feel quite dated at times, and it even changes the campaign missions to be truer to Lord of the Rings lore. Editors Recommendations