Post-Cutoff

Interview

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

David SenraYouTube184,262 views as of 13 October 2026

Watch on YouTubePlay loads YouTube’s player from youtube-nocookie.com.

Why it is here

72-minute interview with Meta’s chief AI officer Alexandr Wang, published the day before Muse Wonder (~184k views by Oct 13). Per the description: the seven-month effort to build Muse, why Meta chose personal agents over coding, how Zuckerberg recruited him after the Llama 4 setback, and the roles of Zuckerberg and Nat Friedman. The YouTube search page showed the title as ‘Creating Muse, Building Meta’s AI Lab & How to Win the AI Race’. Found on Oct 13 through a relevance-sorted search; the launch-day pass had missed it.

Description

Description written by Gemini from the videoGemini 3.8 Flash, 13 October 2026

Summary
In this interview episode, podcast host David Senra interviews Alexandr Wang, leader of Meta Superintelligence Labs (MSL), on the origin, development, and explosive viral growth of Meta’s personal AI agent, Muse. Wang details the transition from Scale AI to Meta, early inspiration from OpenClaw, the internal design philosophy behind Muse, and MSL’s flat organizational culture.

What is shown

  • [00:03] Interview opening: David Senra asks Alexandr Wang for the full history behind the creation of Muse.
  • [01:05] Wang discusses the founding of Meta Superintelligence Labs (MSL) and Mark Zuckerberg’s internal memo on “Personal Superintelligence” in June 2025.
  • [02:00] Discussion of Nat Friedman experimenting with OpenClaw in early 2026, including prompts for psychological probing and tracking personal habits.
  • [05:40] Development of the first Muse prototype in February 2026 within two weeks, featuring the “Jolly” mascot/bot and “Muse Charm”.
  • [07:05] Discussion of Mark Zuckerberg testing Muse early for home use and MMA training feedback.
  • [14:35] Sponsor segment: Ramp advertisement.
  • [15:38] Sponsor segment: Deel advertisement.
  • [16:51] Senra and Wang discuss Jeremy Stern’s 15,000-word profile of Mark Zuckerberg in Colossus, examining Zuckerberg’s relentless execution style.
  • [22:20] Discussion of Muse Spark 1.3 and the engineering rigor required across hundreds of behavioral evaluation metrics.
  • [24:45] Wang explains the team structure: maintaining a focused group of under 200 engineers and researchers.
  • [27:30] Discussion of viral marketing on X, meme culture, and the “Jolly” mascot meme formats driving organic adoption.
  • [36:00] Sponsor segment: AppLovin advertisement.
  • [37:25] Senra recounts a friend running a $100B company who used Muse to autonomously recover an ID left behind on a private Flexjet flight.
  • [42:20] Wang discusses the philosophical mission of personal AI agents: restoring adult agency and childhood ambition.
  • [52:20] The backstory of Meta’s deal with Scale AI: Zuckerberg’s outreach, Meta acquiring 49% of Scale, and Wang joining to lead MSL after Llama 4 underperformed.
  • [57:00] Discussion of Napoleon’s management philosophy via The Mind of Napoleon and Zuckerberg’s all-in commitment.
  • [62:10] Wang reflects on his 19th birthday party during his Hudson River Trading (HRT) winter internship, attended by future tech founders (Scott Wu, Jeffrey Yan, Jesse Zhang, Vicki Ye).
  • [68:10] Wang contrasts “diamond mining” (power-law breakthrough research) versus “building skyscrapers” (operational efficiency and data annotation).

Claims & numbers

  • Wang states that Meta Superintelligence Labs (MSL) was initiated following a June 2025 memo on “Personal Superintelligence” co-developed with Nat Friedman and Mark Zuckerberg.
  • Wang notes that the first functional Muse prototype was developed in 1–2 weeks in February 2026.
  • Wang states that the combined model and product team that built Muse was kept under 200 people.
  • Wang notes that Meta’s product ecosystem spans 3.5 billion daily active users across Instagram, WhatsApp, and Facebook.
  • Wang references Muse Spark 1.3 as a key checkpoint that achieved full “green” evaluation status across 100+ tracked behavior rows.
  • Senra and Wang note Jeremy Stern’s profile of Zuckerberg in Colossus was roughly 15,000 words long.
  • Wang states that Meta owns 49% of Scale AI as part of the transaction where he and key team members transitioned to lead MSL.
  • Wang mentions that quantitative trading firm Hudson River Trading (HRT) generated $11 billion in revenue and roughly $8 billion in profit in Q2 2026 with fewer than 1,000 employees.
  • Wang states he has over 200 direct reports within MSL’s “TBD” core group to enforce an anti-bureaucratic, flat organization.

Notable quotes

  • [19:05] David Senra: “Mark’s like the fucking... I think the line is ‘Mark is the fucking Terminator.’ And it was like he just does, will not stop...”
  • [41:15] Alexandr Wang: “I think Muse can give everyone the adulthood they dreamed of in childhood.”
  • [51:50] David Senra: “In a world that’s changing all the time, the biggest risk is not taking any.”

Assessment
This is a long-form studio podcast interview produced by David Senra and sponsored by Ramp, Deel, and AppLovin. No live software interfaces or benchmarks are displayed on screen; the video consists entirely of a multi-camera studio dialogue discussing technical strategy, company history, and product development narratives.

Described by gemini-3.8-flash on 2026-10-13 from the video’s audio and frames.

Related

  1. Agents 70 days after the cutoff

    Meta launches Muse, a free consumer personal AI agent