Creating Muse: The Fastest-Growing AI Product Since ChatGPT | Alexandr Wang, Meta

TL;DR Meta launched Muse as the fastest-growing consumer AI app by focusing on reliable personal agents rather than coding tools. The product succeeded through meticulous refinement, a unified design vision, and viral marketing that emphasized user benefits.

Muse targets personal utility instead of developer coding tasks.

• Meta chose "personal super intelligence" over coding agents because it aligns with Meta’s core strength in connecting people.
• The company believes AI should remove daily obstacles so users can focus on their interests and hobbies.
• This strategy leverages Meta’s distribution to three billion daily active users for consumer applications.
• Competitors focused heavily on coding, but Meta identified a gap in reliable personal assistance.

Building a reliable agent required seven months of intense refinement.

• The team created a prototype in two weeks after observing the potential of OpenClaw.
• They defined over one hundred specific behaviors and capabilities needed for an ideal personal agent.
• Engineers used numerical thresholds to ensure every behavior met launch quality standards before release.
• The process involved sanding out model details to make the experience consistent and trustworthy for consumers.

A small team ensured a cohesive and singular product vision.

• Fewer than 200 people worked on both the model and the product interface.
• Leadership avoided adding features from multiple managers to prevent a disjointed user experience.
• Nat Friedman shaped the final product design with a single clear point of view.
• This approach mirrored Steve Jobs’ method of creating products with one distinct perspective.

Viral marketing drove rapid adoption through relatable memes and screenshots.

• Alexandr Wang used his X account to share humorous memes and real user success stories.
• The marketing focused on how the tool helps individuals achieve their goals rather than technical specs.
• Users shared screenshots of the agent completing complex tasks, which fueled word-of-mouth growth.
• This strategy made Meta appear cool and engaged the internet culture directly.

Mark Zuckerberg hired Alexandr Wang to rebuild Meta’s AI lab from scratch.

• The deal gave Meta 49% ownership of Scale AI while allowing Wang to lead the new lab.
• Wang shifted from being a bottleneck founder at Scale to a coach enabling top talent at Meta.
• He hired brilliant researchers directly to create a flat, anti-bureaucratic organization structure.
• The goal was to give exceptional engineers the resources and freedom to explore unknown possibilities.