Latest news with #DeepMind
Yahoo
6 hours ago
- Business
- Yahoo
Odyssey's new AI model streams 3D interactive worlds
Odyssey, a startup founded by self-driving pioneers Oliver Cameron and Jeff Hawke, has developed an AI model that lets users "interact" with streaming video. Available on the web in an "early demo," the model generates and streams video frames every 40 milliseconds. Via basic controls, viewers can explore areas within a video, similar to a 3D-rendered video game. "Given the current state of the world, an incoming action, and a history of states and actions, the model attempts to predict the next state of the world," explains Odyssey in a blog post. "Powering this is a new world model, demonstrating capabilities like generating pixels that feel realistic, maintaining spatial consistency, learning actions from video, and outputting coherent video streams for 5 minutes or more." A number of startups and big tech companies are chasing after world models, including DeepMind, influential AI researcher Fei-Fei Lee's World Labs, Microsoft, and Decart. They believe that world models could one day be used to create interactive media, such as games and movies, and run realistic simulations like training environments for robots. But creatives have mixed feelings about the tech. A recent Wired investigation found that game studios like Activision Blizzard, which has laid off scores of workers, are using AI to cut corners and combat attrition. And a 2024 study commissioned by the Animation Guild, a union representing Hollywood animators and cartoonists, estimated that over 100,000 U.S.-based film, television, and animation jobs will be disrupted by AI in the coming months. For its part, Odyssey is pledging to collaborate with creative professionals — not replace them. "Interactive video [...] opens the door to entirely new forms of entertainment, where stories can be generated and explored on demand, free from the constraints and costs of traditional production," writes the company in its blog post. "Over time, we believe everything that is video today — entertainment, ads, education, training, travel, and more — will evolve into interactive video, all powered by Odyssey." Odyssey's demo is a bit rough around the edges, which the company acknowledges in its post. The environments the model generates are blurry and distorted, and unstable in the sense that their layouts don't always remain the same. Walk forward in one direction for a while or turn around, and the surroundings might suddenly look different. But the company's promising to rapidly improve upon the model, which can currently stream video at up to 30 frames per second from clusters of Nvidia H100 GPUs at the cost of $1-$2 per "user-hour." "Looking ahead, we're researching richer world representations that capture dynamics far more faithfully, while increasing temporal stability and persistent state," writes Odyssey in its post. "In parallel, we're expanding the action space from motion to world interaction, learning open actions from large-scale video." Odyssey is taking a different approach than many AI labs in the world modeling space. It designed a 360-degree, backpack-mounted camera system to capture real-world landscapes, which Odyssey thinks can serve as a basis for higher-quality models than models trained solely on publicly available data. To date, Odyssey has raised $27 million from investors including EQT Ventures, GV, and Air Street Capital. Ed Catmull, one of the co-founders of Pixar and former president of Walt Disney Animation Studios, is on the startup's board of directors. Last December, Odyssey said it was working on software that allows creators to load scenes generated by its models into tools such as Unreal Engine, Blender, and Adobe After Effects so that they can be hand-edited. This article originally appeared on TechCrunch at Sign in to access your portfolio
Yahoo
7 hours ago
- Business
- Yahoo
Odyssey's new AI model streams 3D interactive worlds
Odyssey, a startup founded by self-driving pioneers Oliver Cameron and Jeff Hawke, has developed an AI model that lets users "interact" with streaming video. Available on the web in an "early demo," the model generates and streams video frames every 40 milliseconds. Via basic controls, viewers can explore areas within a video, similar to a 3D-rendered video game. "Given the current state of the world, an incoming action, and a history of states and actions, the model attempts to predict the next state of the world," explains Odyssey in a blog post. "Powering this is a new world model, demonstrating capabilities like generating pixels that feel realistic, maintaining spatial consistency, learning actions from video, and outputting coherent video streams for 5 minutes or more." A number of startups and big tech companies are chasing after world models, including DeepMind, influential AI researcher Fei-Fei Lee's World Labs, Microsoft, and Decart. They believe that world models could one day be used to create interactive media, such as games and movies, and run realistic simulations like training environments for robots. But creatives have mixed feelings about the tech. A recent Wired investigation found that game studios like Activision Blizzard, which has laid off scores of workers, are using AI to cut corners and combat attrition. And a 2024 study commissioned by the Animation Guild, a union representing Hollywood animators and cartoonists, estimated that over 100,000 U.S.-based film, television, and animation jobs will be disrupted by AI in the coming months. For its part, Odyssey is pledging to collaborate with creative professionals — not replace them. "Interactive video [...] opens the door to entirely new forms of entertainment, where stories can be generated and explored on demand, free from the constraints and costs of traditional production," writes the company in its blog post. "Over time, we believe everything that is video today — entertainment, ads, education, training, travel, and more — will evolve into interactive video, all powered by Odyssey." Odyssey's demo is a bit rough around the edges, which the company acknowledges in its post. The environments the model generates are blurry and distorted, and unstable in the sense that their layouts don't always remain the same. Walk forward in one direction for a while or turn around, and the surroundings might suddenly look different. But the company's promising to rapidly improve upon the model, which can currently stream video at up to 30 frames per second from clusters of Nvidia H100 GPUs at the cost of $1-$2 per "user-hour." "Looking ahead, we're researching richer world representations that capture dynamics far more faithfully, while increasing temporal stability and persistent state," writes Odyssey in its post. "In parallel, we're expanding the action space from motion to world interaction, learning open actions from large-scale video." Odyssey is taking a different approach than many AI labs in the world modeling space. It designed a 360-degree, backpack-mounted camera system to capture real-world landscapes, which Odyssey thinks can serve as a basis for higher-quality models than models trained solely on publicly available data. To date, Odyssey has raised $27 million from investors including EQT Ventures, GV, and Air Street Capital. Ed Catmull, one of the co-founders of Pixar and former president of Walt Disney Animation Studios, is on the startup's board of directors. Last December, Odyssey said it was working on software that allows creators to load scenes generated by its models into tools such as Unreal Engine, Blender, and Adobe After Effects so that they can be hand-edited. Error while retrieving data Sign in to access your portfolio Error while retrieving data
The Verge
10 hours ago
- Business
- The Verge
You can now try interactive AI worlds backed by Pixar's cofounder
AI companies have recently been experimenting with interactive, AI-generated worlds. There's an AI-generated version of Quake. An AI-generated Minecraft. Google DeepMind is also building a team to develop models that ' simulate the world.' Now, an AI startup backed by Pixar cofounder Edwin Catmull is trying to put its own spin on the idea — something it calls 'interactive video,' which it's letting people experience as part of a research preview that's available today. The startup, called Odyssey, describes interactive video on its website as 'video you can both watch and interact with, imagined entirely by AI in real-time.' The idea is that you can engage with the video in some way — think a first-person video game but in environments that actually look like the real world instead of one made of polygons. Odyssey hypes it up to be an 'early version of the Holodeck,' though it acknowledges that 'the experience today feels like exploring a glitchy dream — raw, unstable, but undeniably new.' In motion, Odyssey's interactive videos feel like walking through a blurry version of Google Street View. You can walk around the startup's real-time generated worlds using the WASD keys as though it were a game. There are a handful of different worlds you can switch between, like a wooded area with a cabin, a shopping mall, and a parking lot in front of a large building. They're a little different each time, since the system is regenerating what's in your vision each time. But the picture quality is generally pretty fuzzy. For now, you only have two and a half minutes to explore the preview before it stops, but you can reload and hop back in if you'd like. Odyssey says it's using clusters of H100 GPUs in the US and Europe to generate the interactive videos. 'Using that input and frame history, the model then generates what it thinks the next frame should be, streaming it back to you in real-time,' the company writes on its website, adding that process can happen in 'as little as' 40 milliseconds. The current preview isn't going to replace Fortnite anytime soon. Objects only sometimes have collision; in one instance, I was stopped by a fence, but when I tried to walk through a large house, I clipped right through it. In another run, I walked down some stairs only to watch the doorway I was heading toward turn into a brick wall. The preview also acts strangely when you're standing still; I did one full instance where I didn't touch the controls at all, and the model slowly kept turning me left and inched me closer to a wall. In an interview with The Verge, Catmull, who sits on Odyssey's board, couldn't give me a specific answer for when the image quality might get better. But he says that Odyssey is on 'the leading edge' of the work that's being done and that 'they participate in this broader community, so the information about how to do this keeps improving.' He acknowledges that the images are still noisy, but he says that the bulk of the noise, like textures on a building, are 'exactly the kind of thing that applying neural network filters to' is meant to solve. It's no Holodeck yet It's not a great video game, despite how entertaining the quirks and issues can be. And I don't think this is going to replace movies for a while, either; the way the world morphs and changes in unexpected ways is just too distracting, and I think knowing that what you're watching won't melt in front of you is a key part of a good film. It's not even a good merging of the two mediums — yet. While messing around with the preview, you can see that there may be something interesting here. With the speed at which AI tools are evolving, it's not too hard to imagine a version of this that doesn't have quite so many issues. But it's no Holodeck yet, and there's quite a ways to go if AI video is going to get there.
Los Angeles Times
12 hours ago
- Entertainment
- Los Angeles Times
Hollywood isn't ready for AI. These people are diving in anyway
When filmmakers say they're experimenting with artificial intelligence, that news is typically received online as if they had just declared their allegiance to Skynet. And so it was when Darren Aronofsky — director of button-pushing movies including 'The Whale' and 'Black Swan' — last week announced a partnership with Google AI arm DeepMind to use the tech giant's capabilities in storytelling. Aronofsky's AI-focused studio Primordial Soup is producing three short movies from emerging filmmakers using Google tools, including the text-to-video model Veo. The first film, 'Ancestra,' directed by Eliza McNitt, will premiere at the Tribeca Festival on June 13, the Mountain View-based search giant said. Google's promotional materials take pains to show that 'Ancestra' is a live-action film made by humans and with real actors, though it's bolstered with effects and imagery — including a tiny baby holding a mother's finger — that were created with AI. The partnership was touted during Google's I/O developer event, where the company showed off the new Veo 3, which allows users to create videos that include sound effects, ambient noise and speech (a step up from OpenAI-owned competitor, Sora). The company also introduced its new Flow film creation tool, essentially editing software using Google AI functions. Google's push to court creative types coincides with a separate initiative to help AI technology overcome its massive public relations problem. As my colleague Wendy Lee wrote recently, the company is working with filmmakers including Sean Douglas and his famous father Michael Keaton to create shorts that aren't made with AI, but instead portray the technology in a less apocalyptic light than Hollywood is used to. Simply put, much of the public sees AI as a foe that will steal jobs, rip off your intellectual property, ruin your childhood, destroy the environment and possibly kill us all, like in 'The Terminator,' '2001: A Space Odyssey' and the most recent 'Mission: Impossible' movies. And Google, which is making a big bet by investing in AI, has a lot riding on changing that perception. There's a ways to go, including in the entertainment industry. Despite the allure of cost-savings, traditional studios haven't exactly dived headfirst into the AI revolution. They're worried about the legal implications of using models trained on troves of copyrighted material, and they don't want to anger the entertainment worker unions, which went on strike partly over AI fears just a couple years ago. The New York Times and others have sued OpenAI and its investor Microsoft, alleging copyright theft. Tech giants claim they are protected by 'fair use.' AI-curious studios are walking into a wild, uncharted legal landscape because of the amount of copyrighted material being mined to teach the models, said Dan Neely, co-founder of startup Vermillio, which helps companies and individuals protect their intellectual property. 'The major studios and most people are going to be challenged using this product when it comes to the output content that you can and cannot use or own,' Neely said by phone. 'Given that it contains vast quantities of copyrighted material, and you can get it to replicate that stuff pretty easily, that creates chaos for someone who's creating with it.' But while the legacy entertainment business remains largely skeptical of AI, many newer, digitally-native studios and creators are embracing it, whether their goals are to become the next Pixar or the next Mr. Beast. The New York Times recently profiled the animation startup Toonstar, which says it uses AI throughout its production process, including when sharpening storylines and lip-syncing. John Attanasio, a Toonstar founder, told the paper that leaning into the tech would make animation '80 percent faster and 90 percent cheaper than industry norms.' Jeffrey Katzenberg, the former leader of DreamWorks Animation, has given a similar estimate of the potential cost-savings for Hollywood cartoons. Anyone working in the traditional computer animation business would have to gulp at those projections, whether they turn out to be accurate or not. U.S. animation jobs have already been hammered by outsourcing. Now here comes automation to finish the job. (Disney's animated features cost well over $100 million to produce because they're made by real-life animators in America.) Proponents of AI will sometimes argue that the new technology isn't a replacement for human workers, but rather a tool to enhance creativity. Some are more blunt: Stop worrying about these jobs and embrace the future of uninhibited creation. For obvious reasons, workers are reluctant to buy into that line of thinking. More broadly, it's still unclear whether all the spending on the AI arms race will ultimately be worth the cost. Goldman Sachs, in a 2024 report, estimated that companies would invest $1 trillion in AI infrastructure — including data centers, chips and the power grid — in the coming years. But that same report raised questions about AI's ultimate utility. To be worth the gargantuan investment, the technology would have to be capable of solving far more complex problems than it does now, said one Goldman analyst in the report. In recent weeks, the flaws in the technology have crossed over into absurd territory: For example, by generating a summer reading list of fake books and legal documents polluted with serious errors and fabrications. Big spending and experimentation doesn't always pan out. Look at virtual reality, the metaverse and the blockchain. But some entertainment companies are experimenting with the tools and finding applications. Meta has partnered with horror studio Blumhouse and James Cameron's venture Lightstorm Vision on AI-related initiatives. AI firm Runway is working with Lionsgate. At a time when the movie industry is troubled in part due to the high cost of special effects, production companies are motivated to stay on top of advancing tech. One of the most common arguments in favor of giving in to AI is that the technology will unshackle the next generation of creative minds. Some AI-enhanced content is promising. But so far AI video tools have produced a remarkable amount of content that looks the same, with its oddly dreamlike sheen of unreality. That's partly because the models are trained on color-corrected imagery available on the open internet or on YouTube. Licensing from the studios could help with that problem. The idea of democratizing filmmaking through AI may sound good in theory. However, there are countless examples in movie history — including 'Star Wars' and 'Jaws' — of how having physical and budgetary restrictions are actually good for art, however painful and frustrating they may have been during production. Even within the universe of AI-assisted material, the quality will vary dramatically depending on the talent and skill of people using it. 'Ultimately, it's really hard to tell good stories,' Neely said. 'The creativity that defines what you prompt the machine to do is still human genius — the best will rise to the top.' Like other innovations, the technology will improve with time, as the new Google tools show. Both Veo 3 and Flow showcase how AI is becoming better and easier to use, though they are still not quite mass-market products. For its highest tier, Google is charging $250 a month for its suite of tools. Maybe the next Spielberg will find their way through AI-assisted video, published for free on YouTube. Perhaps Sora and Veo will have a moment that propels them to mainstream acceptance in filmmaking, as 'The Jazz Singer' did for talkies. But those milestones still feel a long way off. The Memorial Day weekend box office achieved record revenue (not adjusting for inflation) of $329.8 million in the U.S. and Canada, thanks to the popularity of Walt Disney Co.'s 'Lilo & Stitch' and Paramount's 'Mission: Impossible — The Final Reckoning.' Disney's live-action remake generated $183 million in domestic ticket sales, exceeding pre-release analyst expectations, while the latest Tom Cruise superspy spectacle opened with $77 million. The weekend was a continuation of a strong spring rebound for theaters. Revenue so far this year is now up 22% versus 2024, according to Comscore. This doesn't mean the movie business is saved, but it does show that having a mix of different kinds of movies for multiple audiences is healthy for cinemas. Upcoming releases include 'Karate Kid: Legends,' 'Ballerina,' 'How to Train Your Dragon' and a Pixar original, 'Elio.' 'Lilo & Stitch' is particularly notable, coming after Disney's previous live-action redo, 'Snow White,' bombed in theaters. While Snow White has an important place in Disney history, Stitch — the chaotic blue alien — has quietly become a hugely important character for the company, driving enormous merchandise sales over the years. The 2002 original wasn't a huge blockbuster, coming during an awkward era for Walt Disney Animation, but the remake certainly is. Watch: Prepping for the new 'Naked Gun' by rewatching the classic and reliving the perfect Twitter meme. Listen: My favorite episode of 'Blank Check with Griffin & David' in a long time — covering Steven Spielberg's 'Hook' with Lin-Manuel Miranda.
India Today
18 hours ago
- Business
- India Today
New Google AI tool translates sign language into text, currently in testing phase with launch by year-end
Sign language is essential for many people who have speech impairment. They use it to communicate with people around them but among the regular not many understand it. Now, AI is going to help here as well. Google is working on a AI model called SignGemma that will translate sign language into text. The company says this is its most capable artificial intelligence model to date, designed to translate sign language into spoken text. This new AI model is currently in its testing phase, and is slated for public launch by the end of the first unveiled SignGemma during the keynote at Google I/O, where Gemma Product Manager Gus Martins described it as the company's 'most capable sign language understanding model ever.' Martins noted that, unlike previous attempts at sign language translation, SignGemma stands out for its open model approach and its focus on delivering accurate, real-time translations to users. While the tool is trained to handle various sign languages, Google says the model currently performs best with American Sign Language (ASL) and English.'We're thrilled to announce SignGemma, our groundbreaking open model for sign language understanding, set for release later this year,' Martins said. 'It's the most capable sign language understanding model ever, and we can't wait for developers and Deaf and hard-of-hearing communities to take this foundation and build with it.' Google highlighted that with this tool, the company aims to bridge communication gaps for millions of Deaf and hard-of-hearing individuals to ensure the tool is both effective and respectful of its user base, Google is taking a collaborative approach to its development. The company has extended an open invitation to developers, researchers, and members of the global Deaf and Hard of Hearing communities to participate in early testing and provide feedback."We're thrilled to announce SignGemma, our groundbreaking open model for sign language understanding," reads the official post from DeepMind on X. "Your unique experiences, insights, and needs are crucial as we prepare for launch and beyond, to make SignGemma as useful and impactful as possible."The introduction of SignGemma comes at a time when Google is heavily focused on expanding its AI portfolio. At Google I/O 2025, accessibility took centre stage with the announcement of several new AI-powered features designed to make technology more inclusive for everyone. One of the highlights was the expansion of Gemini AI's integration with Android's TalkBack, which will now provide users AI-generated descriptions for images and allow them to ask follow-up questions about what's on their screen. Google has also introduced updates to Chrome, including automatic Optical Character Recognition (OCR) for scanned PDFs, enabling screen reader users to access, search, and interact with text in documents that were previously inaccessible. For students, on Chromebooks a new accessibility tool called Face Control allows users to control their device with facial gestures and head movements.
