Nvidia released DreamDojo, a new AI system designed to teach robots how to interact with the physical world by watching tens of thousands of hours of human video, a development that could significantly reduce the time and cost required to train the next generation of humanoid machines, according to VentureBeat. The research, published this month, introduces what the team calls "the first robot world model of its kind that demonstrates strong generalization to diverse objects and environments after post-training." Meanwhile, the tech world also saw the launch of Even Realities' G2 smart glasses, which, according to Wired, boast a larger, sharper display and a more comfortable frame than its predecessor, the G1. However, the software still needs some work.
DreamDojo, developed by a team of researchers led by Nvidia, utilizes a large-scale video dataset to train robots. The project involved collaborators from UC Berkeley, Stanford, the University of Texas at Austin, and other institutions, VentureBeat reported. The system aims to provide robots with a "world model," allowing them to generalize and adapt to various objects and environments.
Even Realities launched the G1 smart glasses 19 months prior, and the market has evolved rapidly since then, according to Wired. The G2 smart glasses feature a titanium and magnesium frame, smart AI integration, and improved battery life. They also support a wide range of prescription lenses. The optional R1 smart ring is available to simplify controls. However, software stability remains an issue, and the display can be difficult to read in bright light. The smart ring is an extra $249.
In other tech news, the online hangout Moltbook, populated by AI agents interacting with one another, was compared to Pokémon by MIT Technology Review. The platform, which showed AI systems doing useful things for their creators, reminded the publication's senior editor for AI, Will Douglas Heaven, of the popular game. The platform was also flooded with crypto scams, and many of the posts were written by people.
MIT Technology Review also announced the launch of a new weekly newsletter series, "Making AI Work," which will explore how generative AI is being used and deployed across sectors. Each edition will begin with a case study and then delve deeper into the AI tools being used.
Discussion
AI Experts & Community
Be the first to comment