Google outlines latest step towards creating artificial general intelligence
Briefly

Google has introduced the Genie 3 world model, facilitating AI interaction with realistic simulations. This model is essential for training robots and autonomous vehicles in environments like warehouses. DeepMind asserts that such world models are pivotal for attaining artificial general intelligence (AGI), where systems would perform tasks comparable to humans. Although Genie 3 shows promise, it is not yet publicly available and has various limitations. This technology is intended to aid in simulating experiences and tasks effectively, reinforcing the move towards AGI in a competitive AI landscape.
Google's Genie 3 world model advances AI interaction with realistic simulations, vital for training robots and autonomous vehicles, aiming towards artificial general intelligence.
DeepMind emphasizes that world models are essential for achieving AGI, enabling AI systems to perform most human tasks autonomously rather than just specific functions.
Genie 3 could create simulations for various training experiences, such as skiing or exploring environments, by generating scenarios from text prompts and modifying them quickly.
Despite promising insights, Google notes Genie 3 remains unreleased for public use and still possesses numerous limitations that must be addressed.
Read at www.theguardian.com
[
|
]