Meta's Muse AI Agent Bets on Consumers as Rivals Chase Enterprise

On the podcast, reporters Kirsten Korosec, Sean O'Kane and Anthony Ha said Meta's AI announcements appeared to overshadow a week of new model launches from OpenAI and Anthropic. Korosec said the consumer approach may suit Meta, which has an established record of embedding itself in everyday life through Facebook, Instagram and WhatsApp. "I've never really thought of them as an enterprise product anyway," she said.

Ha said the direction felt "head spinning" given the industry's wider shift toward enterprise. He said OpenAI and Anthropic, which are planning to go public — OpenAI possibly this year or next — are under pressure to generate money at an unprecedented scale, and see enterprise as the way to do it. Meta, he said, may see a different opening.

O'Kane said he had used Muse, which had been available for about two weeks at the time of the recording. He described it as roughly Meta's ground-up version of an on-phone app comparable to OpenClaw, the agent tool Meta acquired and integrated months earlier. The app works as a "relatively good chatbot" interface and suggests tasks it can perform, he said.

One suggestion was to scan for unclaimed funds. O'Kane said Muse found some in his name on his first day of use and that a check is on its way to him. But he described the feature as more "a party-trick type thing" than something that would drive ongoing usage. He also raised the question of whether users would trust Meta's AI with sensitive information. "Meta's business is to sell you ads," he said. "And yes, they'll make the argument that the more they know about you, the more accurate and interesting the ads will be — wake me up when we get to that fever dream."

Meta also showed a Tamagotchi-style AI device at Connect that the company insists is for adults only.