Meta releases its most powerful open-source AI model, Zuckerberg challenges OpenAI and Anthropic

Meta announced the release of the weights of its most powerful AI model and launched a new series of models designed specifically for consumer-grade devices. This move is aimed at competing with leading labs such as OpenAI and Anthropic, while also demonstrating to investors that its massive investment in AI is yielding results.

On Monday, August 10, Meta CEO Mark Zuckerberg announced in an Instagram video thatthe company will open-source its latest AI model, Muse Spark 1.2, allowing the public to download and use it. At the same time, the company will launch the open-source model series Muse Glimmer, designed specifically for laptops.

Zuckerberg's high-profile announcement aims to demonstrate to the market that the Meta Super Intelligence Lab, established last year, has made substantial progress and to endorse the expected $145 billion capital expenditure for this year.

After the announcement, Meta's stock price rose as much as 2.6% on Monday. Previously, due to continued investor scrutiny of the company's spending plans and AI competitiveness, the stock had fallen about 10% year to date.

01

Betting on Open Source: A Differentiated Strategy to Compete with Closed Giants

Meta's commitment to open source has a clear strategic logic.

In a 6,500-word AI paper released on the same day, Zuckerberg positioned Meta's open source approach as a key force against closed AI ecosystems and called on Washington to support US open source efforts.

Zuckerberg also issued a clear call regarding US AI policy. He wrote:

Foreign labs currently hold a number of advantages because US labs must comply with more additional restrictions on training data. If the US wants open source models to maintain a long-term lead, policy must reduce these obstacles.

He also stated:

I do not believe restricting access to foreign open source models is an effective solution. Our goal should be for US open source models to become the best choice globally, which requires removing the obstacles that make it harder for US open source to compete.

Counterpoint Research co-founder Neil Shah commented:

If western tech giants only build walled gardens, developers and businesses will naturally turn to open-weight models. Most of Meta's American competitors use proprietary approaches, but the market has a strong demand for non-closed open-source models,Meta is perfectly positioned to fill this gap.

Muse Glimmer focuses on edge-side deployment, making it distinct from the mainstream cloud-based AI processing approach. Neil Shah noted:

Deploying lightweight agent models like Muse Glimmer directly on PCs and mobile devices can bypass the costs of cloud computing power, creating a competitive advantage on the user side versus Google, Microsoft, and others.

02

Aimed at OpenAI and Anthropic

While outlining his open-source vision, Zuckerberg warned about the dangers of AI power concentration—a statement widely interpreted as a veiled critique of OpenAI and Anthropic.

He wrote:

The idea that AI is so dangerous that the only safe path is extreme concentration of power is fundamentally flawed. Surprisingly, many AI developers use doomsday rhetoric. I don't understand why someone who believes AI will eliminate most jobs and human value is so eager to build such a future.

Anthropic CEO Dario Amodei and OpenAI CEO Sam Altman have both previously warned about the impact of AI on employment; Altman has recently softened some of his related statements.

Zuckerberg, on the other hand, put forward a completely different vision:

We should widely distribute superintelligence, not centralize it, and empower everyone to harness its capabilities. Each person will have a super-intelligent personal agent that deeply understands themselves, their goals, and everything they care about.

If you enjoyed this article, please click “Watching”