A Swedish AI startup, IntuiCell, unveiled a robot dog named Luna on Wednesday, equipped with a groundbreaking digital nervous system that enables it to learn and adapt similarly to humans and animals.
Luna represents one of the earliest examples of physical agentic AI, capable of making decisions and taking actions to achieve specific objectives, rather than being limited to narrow tasks or content generation, mirroring the learning abilities of a real dog.
Instead of relying on a generative AI model and extensive datasets, IntuiCell plans to employ a dog trainer to teach Luna how to walk, leveraging the interaction and processing of neurons within its digital nervous system.
CEO and co-founder Viktor Luthman stated in an interview, "We've developed the first software that enables any machine to learn in the same way humans and animals do."
He emphasized that Luna's learning process requires no pre-training, offline simulations, or massive data centers—just a nervous system that facilitates machine learning.
Luthman believes the robot dog's capabilities could pave the way for advancements in human-like robots operating in various environments.
He added, "Our next focus is humanoid robotics and autonomous systems for unpredictable settings like space exploration, deep-sea missions, or disaster response."
For instance, intelligent machines like Luna could be deployed to Mars to construct habitats for future human settlers, adapting to unforeseen challenges in real-time without prior training, Luthman explained.
Comments
Post a Comment