From CNET: This seems to be the year of AI agents, and Meta is diving into the game. The company on Tuesday announced a new agent program called Muse, which CEO Mark Zuckerberg describes as a “personal agent” that understands your habits and goals. It’s available now in the US.
Agentic AI is a type of generative AI based on a large language model where agents, or bots, are meant to take actions autonomously on your behalf. Meta’s Muse, for example, can do online shopping (with your permission), create documents, fill out forms and respond to emails.
Meta says that Muse is a more general-purpose agent, not limited to specific tasks like coding. (If you want agentic coding help, you can use Muse Code, also recently released by the company.) You have to give Muse permission to access your apps and data, which you can revoke at any time. Like a lot of agentic AI tools, the more info you give it, the more personalized your results will be.
Muse runs “on its own dedicated computer in the cloud,” according to a Meta blog post, in what the company calls a “virtual machine.” The tool also has a separate Sentinel agent — you can think of it as a virtual traffic cop, giving Muse a red or green light before letting it out onto the internet to complete tasks.
View: Full Article
