Meta is rolling out upgraded AI models to its Muse and Meta AI assistants that are built to remember what matters from past chats and keep going when a task hits a snag.
The company announced the update Friday. The new models were trained using Muse Wonder, the company’s most capable model series, which it introduced a few weeks ago. They add two features called Active Memory and Active Focus.
Remembering what matters
Active Memory is meant to help the assistant recall earlier information when it becomes useful. Meta says it trained the models to tell temporary information from permanent, and what is still current from what has gone stale. In Meta’s own tests, the new Muse model used critical context correctly in 90% of tasks, up from 76% for the previous model. Meta also says it cut the forgetting of critical information by 2.4 times.
Sticking with the task
Active Focus is meant to keep Muse working toward a goal when something unexpected gets in the way. Meta says the models try harder, recover, or find another route. In its own tests, Muse completed everyday tasks with unexpected problems 67% better than before. The models can also juggle several jobs, decide when to work alone or with you, and take a correction in the middle of a task without getting confused.
Meta AI gets the same upgrade
In Meta AI, the company says the same changes make it a better partner for thinking things through. It can break a big problem into steps, ask sharper follow-up questions, and keep up when you change direction mid-conversation. On Meta’s Personal Superintelligence Index, the new Muse model scored 47%, up from 38%.
In plain words: Meta says its assistants should now remember the useful parts of your past chats and not give up when a task gets tricky. The scores come from Meta’s own tests.
JBizNews Desk | Menlo Park, California
© JBizNews.com. All rights reserved. Unauthorized reproduction or redistribution is strictly prohibited.
