Meta (META.US “) Muse or reshape the AI competitive landscape! Analyst: Personal intelligence moats far surpass chatbots, OpenAI and Anthropic may be under pressure

The Zhitong Finance App learned that Stratechery analyst Ben Thompson said that the strategic threat posed by Meta (META.US)'s Muse to cutting-edge AI laboratories such as OpenAI and Anthropic may be greater than previously estimated.

Thompson called Muse “the best and easiest personal intelligence product I've ever used, with a clear advantage.” His core observation is that Meta's Muse Spark 1.3 model isn't the most advanced, but it's enough to drive a compelling AI agent.

“This is exactly a bearish sign,” Thompson wrote. He believes that the AI model's capabilities may be mature enough to enable the company to create consumer products with user stickiness without leading the model competition.

Why Muse is important

Thompson believes that the moat of personal intelligence may be much stronger than chatbots. Once users feed personal information to the smart body and incorporate it into their daily lives, it becomes much more difficult to switch to other services.

“The model ability is good enough to create an attractive product—one that really has a moat,” he wrote.

This may challenge the strategies of cutting-edge laboratories: they have been investing heavily in building increasingly powerful models, while moving towards mastering the user interface and the applications built on it.

Thompson also believes that the AI industry is moving towards a stratification of models and “harnesses” (driving frameworks) — software layers that turn models into useful agents. OpenAI's Codex and ChatGPT, and Anthropic's Claude Code are examples of this intelligent integration path, but the programming products they generate can eventually be used with competitors' models.

Smart bodies are becoming the main battleground

This argument came at a time when AI agents became the next major battleground for the industry. Both OpenAI and Anthropic are advancing intelligent tools that can perform multi-step tasks, particularly programming and computer manipulation workflows.

Thompson has previously said that intelligent AI is powerful enough to justify the huge demand for computing infrastructure. However, Muse suggests that value may increasingly be skewed towards the party that owns the end user product rather than the party with the strongest underlying model.

Traders are eyeing Meta's AI layout

Meta's stock price rose more than 11% on Monday, increasing its market value by nearly $200 billion. Previously, the Muse app reached the top of the US iPhone and Android download list.

Muse can handle multi-step tasks, including browsing websites, filling out forms, and performing actions on behalf of users. The launch of the app also sparked a backlash from the platform side, and Amazon has blocked Muse from visiting its website.

A trader said that META's Monday trend “stemmed from early signals that its AI agents are working. A week ago, market pricing was still debating whether AI development would slow down. This is what repricing is expected to look like. This smart body has barely generated revenue, but the market is moving about what it could become.”

Muse could give Meta a deeper understanding of users

Muse may enhance Meta's advertising and recommendation system by providing a far richer understanding of users' interests and behaviors than ever before.

Instagram director Adam Moseri recently stated that Meta's system has historically relied heavily on signals such as “what users like and what people with similar interests are interacting with.” On Lenny's Podcast in July, he said, “Until now, we've really reached the level of detail people have assumed we had for years.”

This situation is likely to change further as users spend more time with personal AI agents like Muse that can remember conversations, preferences, and details of their lives.