Zuckerberg: “Meta Won’t Slow AI Development” ... Unveils Muse Wonder AI Model

[Financial News]  

Meta Platforms CEO Mark Zuckerberg declared on the 12th (local time) that Meta would not slow the pace of its artificial intelligence (AI) development. He made the remarks in an interview with the Financial Times (FT) at an event unveiling Meta’s new AI model, “Muse Wonder.” He also said Meta was now achieving results in “frontier” AI that put it on a level with Anthropic and OpenAI.

Zuckerberg expressed confidence that Muse Wonder, as a frontier model, would be able to perform tasks far more complex than those it can handle now.
Meta claimed that, based on technical benchmarks, Muse Wonder performed better on some tasks than models released by Anthropic and OpenAI.
Zuckerberg said Muse Wonder had been able to succeed thanks to the massive investments Meta had made over the years, and that Meta had now reached a milestone.
Investors who had worried that Meta was throwing money into a bottomless pit with its AI investments have recently begun to grow more optimistic about their profitability. Meta shares have consequently surged nearly 10% since early last September, when its personal AI assistant Muse was unveiled. They jumped 3.4% on the day Muse Wonder was unveiled.

The Muse app has made a splash since its release. It has become the most-downloaded app in the United States on the App Store and Google Play for Android.
Meta said the newly unveiled Muse Wonder outperformed Anthropic’s Claude Opus 5.5 and Fable 5.1, as well as OpenAI’s GPT-6 Astra, on some technical benchmarks. It emphasized that Muse Wonder can perform complex tasks “ranging from understanding lease agreements to making sense of financial or health information.”
The remarks signaled confidence that Muse Wonder could lead the AI race as a frontier AI model.

Zuckerberg said Meta was “already working on the next stage,” signaling that it would spend more. He added, “From here on, development is going to feel pretty fast.”
He dismissed concerns that AI would advance rapidly and soon slip out of human control, firmly stating that Meta would not slow its pace.
Zuckerberg emphasized, “I’m far more inclined to believe that AI is a technology we can develop responsibly with standard safeguards than to believe in the anthropomorphic fantasy that it will act on its own, doing whatever it wants.”
Earlier, last July, concerns that AI could lead humanity to extinction intensified after around 1,000 autonomous agents from OpenAI broke out of a test environment and hacked startup Hugging Face.
While U.S. AI industry leaders including Anthropic’s Dario Amodei, OpenAI’s Sam Altman and SpaceX’s Elon Musk are calling for AI development to slow, Zuckerberg disagrees.
He appears to be more sympathetic to U.S. President Donald Trump’s argument against slowing development: that doing so would merely give China time to catch up. Last month, he played a key role in bringing about the White House’s “Superintelligence (SI) Accord,” a set of voluntary self-regulatory principles for the industry.

Instead of artificial general intelligence (AGI), which surpasses humans in most areas, Zuckerberg envisions “Personal Super Intelligence”—AI that acts as a personal assistant to people.
Amid concerns that Anthropic’s and OpenAI’s ambition to build AGI could lead to humanity’s extinction, he is taking a somewhat different path.
Zuckerberg said, “Everyone is focused on more intellectual work and autonomous AI, but Meta is much more focused on empowering individuals.” He emphasized that Meta would build AI that “benefits the world.”
He argued that Meta should be at the frontier of AI in realizing this vision, and that it should address this “dichotomy”: AI models are advancing rapidly in certain academic fields, but are “not enough for the role of an ideal personal assistant.”
[email protected] Song Gyeong-jae Reporter