Meta Unveils V-JEPA 2: A smarter AI that understands the physical world
Meta has launched V-JEPA 2, an advanced AI model trained on video, designed to help robots and AI systems better understand and predict how the physical world works. This model represents a big step toward Meta's long-term goal of developing Advanced Machine Intelligence (AMI)—AI that can think before it acts.
Humans naturally understand how the world responds to actions—for example, knowing a ball will fall after being tossed, or avoiding people while walking through a crowd. This is because we build mental models of how things move and interact. V-JEPA 2 aims to give AI agents a similar understanding of their surroundings, helping them to observe, reason, and make better decisions before taking action.
These types of AI systems are built using world models, which allow machines to understand, predict, and plan—just like humans do when navigating everyday situations.
Smarter robots through video training
V-JEPA 2 builds on Meta's earlier version of the model released last year, called V-JEPA. This new version offers major improvements in helping machines predict how objects move, how they interact with other objects, and how people handle them in the real world.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Time of India
3 hours ago
- Time of India
Scale AI may lose partnership with Microsoft, Google, ChatGPT-maker OpenAI and Elon Musk's xAI after Facebook's $14 billion investment
Google, the largest customer of AI data-labeling firm Scale AI, plans to end its relationship with the company following news that rival Facebook-parent company Meta is acquiring a 49% stake, a report said. Google had reportedly intended to pay Scale AI around $200 million this year for human-labeled training data, essential for developing advanced AI models like its ChatGPT competitor, Gemini. Citing five sources familiar with the matter, news agency Reuters reported that the move stems from concerns that Meta's significant investment could expose Google's proprietary AI research and roadmap to a key competitor. Reportedly, the search giant has already begun discussions with several of Scale AI's rivals this week, looking to shift this substantial workload. Not only Google, The unease isn't limited to Google. Other major tech companies using Scale AI's services, including Microsoft and Elon Musk's xAI, are also reportedly distancing themselves. The report also says that OpenAI, another significant AI developer, had already scaled back its engagement with Scale AI months ago, though its spending was considerably less than Google's. However, OpenAI's CFO recently stated that the company would continue working with Scale AI as one of its many data vendors. Why these companies are scaling back their involvement in Scale AI The primary concern among companies competing with Meta in cutting-edge AI model development is the risk of their research priorities and technical blueprints being exposed to a direct rival. Customers often share proprietary data and even prototype products with Scale AI's workers for data-labeling services. With Meta now holding a substantial stake, these AI firms fear their chief competitor could gain sensitive business and technical intelligence. What it may mean for Scale AI For Scale AI, this potential loss of Google's business comes as a blow, despite its recent valuation jump to $29 billion following Meta's investment (up from $14 billion pre-deal). Scale AI's core business heavily relies on a few major customers, making it vulnerable to such departures. While Scale AI intends to continue operations, its CEO, Alexandr Wang, along with some employees, are moving to Meta. The bulk of Scale AI's revenue comes from providing human trainers to annotate complex datasets for generative AI model makers, a service that can cost as much as $100 per annotation for sophisticated examples. Its largest revenue stream is from its partnerships with generative AI model makers.
Time of India
5 hours ago
- Time of India
From MIT dropout to AI billionaire: Who is Scale AI's Alexandr Wang?
Alexandr Wang , the CEO and cofounder of Scale AI , has emerged as a significant figure in the artificial intelligence (AI) sector. Born in New Mexico, US, Wang's parents were Chinese immigrants who worked as nuclear physicists at Los Alamos National Laboratory. Before attending college, Wang gained early experience working at Quora. Wang enrolled at the Massachusetts Institute of Technology (MIT) but left after just one year. He then joined startup accelerator Y Combinator. During his time at Y Combinator in 2016 Wang partnered with Lucy Guo, also a Quora alum, to establish Scale AI. Within two years, both Wang and Guo were recognised in Forbes' "30 Under 30" list for enterprise technology. Soon after this recognition, Guo departed from Scale AI 'due to differences in product vision and road map,' according to a report by Forbes. Though other reports indicate that she was removed from her position. Unicorn status Established in 2016, Scale AI focusses on providing large volumes of accurately labelled data essential for training advanced AI systems, including tools like OpenAI's ChatGPT . The company achieved unicorn status in 2019 after securing $100 million in funding from Peter Thiel's Founders Fund. This was followed by another fundraising round that brought in $580 million, valuing the company at $7 billion. At 24, Wang became the world's youngest self-made billionaire. His co-founder, Lucy Guo, later became the youngest self-made woman billionaire, owing to her stake in Scale AI. Discover the stories of your interest Blockchain 5 Stories Cyber-safety 7 Stories Fintech 9 Stories E-comm 9 Stories ML 8 Stories Edtech 6 Stories Recent investment by Meta Recently, tech major Meta invested in Scale AI, reportedly valuing the data labelling startup at over $29 billion. Meta acquired a 49% stake for $15 billion, as reported by Reuters . This move shows Meta's strong push to compete with rivals like OpenAI, Google, and Microsoft in the AI space. As part of this agreement, Wang is set to take a leadership role at a newly established, 50-person research lab at Meta, tasked with creating artificial superintelligence (ASI) -- AI systems designed to surpass human intelligence.
Time of India
6 hours ago
- Time of India
Google rejects app store age verification for online content
American tech heavyweight Google on Friday reiterated its opposition to verifying the age of a device's user through the app stores built into operating systems, calling a proposal from Facebook and Instagram parent Meta "ineffective".Limiting access to age-restricted content online is a live issue in Europe, with France battling pornography sites over its newly-introduced requirement that they check users' is also one of several capitals pressing Brussels to introduce Europe-wide regulations cutting off access to social networks for under-15s over concerns including addiction, cyberbullying and hate speech. Basing age verification on details from a device's app store "would require the sharing of granular age band data with millions of developers... who don't need it", such as producers of uncontroversial apps like flashlights, Google wrote in a blog post. "We have strong concerns about the risks this 'solution' would pose to children," it added. The search giant's Play Store is a part of the Android operating system, by far the most widely-used around the globe. Google said that using app store data to verify ages would also leave major ways people access content online unprotected for the underage, such as desktop computers or shared family devices. Apple -- whose own App Store is loaded on every device running its iOS operating system, such as iPhones and iPads -- has also pushed back against Meta's proposal. "The right place to address the dangers of age-restricted content online is the limited set of websites and apps that host that content," the iPhone maker said in a February document. "Implementing age verification at the operating system or app store level will help ensure that we create an ecosystem that's safe for teens," the Facebook owner's safety chief Antigone Davis told Euronews in February. Meta has sinced launched a campaign for European regulation to require the measure. Europe's Digital Services Act (DSA), which came into force last year, says it is up to platforms like Meta's to verify the age of their users -- not providers of operating systems or app stores. Google said that changing to the latter system -- which has also been pushed by Pornhub parent company Aylo -- would mean "reengineering the protocols that have defined the decentralized web in ways that are hard to fully predict".
