Alexandr Wang's First Meta Interview | Generated by AI

Home 2026.05

Question: What did Alexandr Wang, Meta’s MSL (Meta Superintelligence Labs) lead, say in his recent interview with Ashlee Vance on the Core Memory podcast? (Context: “Alexandr Wang Responds to Everything: LeCun, Manus, ‘My Parents Are Both Chinese’”)

Answer:

Alexandr Wang, Meta’s Chief AI Officer and head of Meta Superintelligence Labs (MSL), gave his first major public interview since joining Meta, speaking with tech journalists Ashlee Vance and Kylie Robison on the Core Memory podcast (released ~May 13, 2026). Here is a comprehensive breakdown of what he covered:

1. His Background & Chinese Heritage

Wang was born in Los Alamos, New Mexico, in January 1997. He is the son of Chinese immigrants who worked as physicists at the Los Alamos National Laboratory. This became relevant context for questions about his identity amid the US-China AI rivalry, as he has been vocal about America needing to “win the AI war” while his parents are both Chinese.

2. Why Meta Was Behind in AI — and What He Did About It

Wang said: “When I got to Meta, it was clear that there needed to be some reset of the efforts, and some rebuild of our AI efforts to get onto the right trajectory. Because ultimately, Llama 4 was not on the same trajectory, and so we were behind the frontier. And so we needed to build a plan that would enable us to have a very, very fast velocity, to be able to both catch up and hopefully exceed where the frontier is.”

Llama 4 was criticized as lagging behind rival models — Meta’s own chief scientist Yann LeCun later admitted benchmarks had been “fudged a little bit” — and the reception damaged Meta’s credibility as a serious frontier lab.

3. Muse Spark — His First Deliverable at Meta

MSL rebuilt the entire stack from scratch — architecture, data pipelines, training infrastructure — and the new pretraining stack is sufficient to achieve the same capabilities with over an order of magnitude less compute than its predecessor, Llama 4 Maverick.

Muse Spark, codenamed “Avocado,” launched in April 2026 and tied for 4th place on major AI benchmarks. It is now being integrated across WhatsApp, Instagram, Facebook, and Meta’s smart glasses.

4. On Manus (the AI Agent Acquisition)

Meta picked up the crucial software building block in late December 2025: Manus, a Singapore-based provider of autonomous AI agents that originally emerged from a Chinese startup. Meta closed the deal at over $2 billion, and as part of the acquisition, all remaining Chinese ties to the company are being severed. Unlike conventional chatbots, the Manus agent handles multi-step tasks largely on its own — from market research and coding to data analysis — and recently reached an annualised revenue run rate of over $125 million.

5. On Meta Researchers Being “Just Money Motivated”

Vance asked Wang about the reputation that Meta’s Superintelligence Lab is composed of top AI researchers who were lured away from rivals thanks to very generous offers, with top AI researchers reportedly receiving $100 million packages. Wang pushed back: “I think it’s like an incorrect assumption to think that the researchers are just money motivated or anything. And for most of them actually, the financial prospects of them staying wherever they were looked very good as well — very, very, very strong as well.”

Wang described this as “one of the larger narrative violations or maybe like differences between external perception and what the day-to-day inside is like actually.”

6. On Yann LeCun

Yann LeCun, Wang’s former colleague, had publicly called Wang “inexperienced” and said he didn’t fully understand AI researchers. On the podcast, Wang said he saw LeCun in India a few weeks after the comment, and that LeCun had congratulated Meta Superintelligence Labs on the launch of Muse Spark. Wang described LeCun as “a notable, very outspoken person” and said “I think everyone always knows what Yann is thinking.”

Notably, Ashlee Vance said they placed the “Yann bait” in front of Wang but he would not take it — meaning Wang refused to escalate the drama or attack LeCun directly.

7. On AI Industry Tensions (Sam Altman & Others)

Host Ashlee Vance said that relationships in the AI community seemed to be getting worse. Wang replied: “Maybe it gets worse, then it gets better.” Wang called any hostilities in the AI industry “unfortunate” and said: “My real hope is that all these animosities subside over time and then people sort of come together and realize we are building this incredibly important technology.”

Vance mentioned he texted Sam Altman — Wang’s former roommate — about the interview. Vance said Altman “did not have flattering things to say.” Wang did not directly comment on his relationship with Altman.

8. On Criticism and Controversies

Wang said: “What’s out there is never always correct. It can be frustrating, but I choose to just channel it into the work that we’re doing.”

9. On Model Welfare (a Surprising Topic)

Wang made the notable claim that “there are ways to measure the sort of subjective experience of the models,” touching on model welfare — the question of whether AI models might have some form of inner experience worth considering.

Summary: This was Wang’s first comprehensive public appearance since joining Meta nearly a year ago. He addressed the full spectrum of controversies — LeCun’s attacks, the Manus acquisition, researcher poaching, Meta’s AI decline under Llama 4, and his own Chinese-American identity — while largely staying diplomatic, letting the results (Muse Spark) do the talking.

References:

Back Donate