A new generation of compact edge-AI computers is moving artificial intelligence out of distant data centres and directly into robots, drones and autonomous machines.
AI News reports that Nvidia’s Jetson Orin Nano 2 is designed to provide improved physical-AI performance within a small computing platform. Devices in this category allow machines to process camera feeds, sensor data and control decisions locally rather than sending every action to the cloud.
Local intelligence reduces delay
A robot operating near people cannot always wait for a round trip to a remote server. Network congestion, poor connectivity or a service outage could interrupt a safety-critical decision. Edge processing lowers latency and can keep essential functions working even when the wider network is unavailable.
It can also strengthen privacy by limiting the amount of raw video or environmental data transmitted elsewhere. The trade-offs include power consumption, thermal management, device security and the difficulty of updating thousands of machines in the field.
MaryChuks analysis: embodiment changes the meaning of intelligence
A language model can pause before answering. A drone approaching an obstacle cannot. Physical AI therefore requires a different relationship between cognition, energy, sensing and time. Compact computing modules are the nervous-system layer that connects machine intelligence to real-world action.
Cloud AI thinks at a distance. Edge AI lets the machine think where consequences happen.
Source: AI News
Discover more from Marychuks.com AI, Psychology, Business & CreativeVerse
Subscribe to get the latest posts sent to your email.
