Adam Mosseri: AI is a tailwind for authenticity

Adam Mosseri, head of Instagram at Meta, discusses how product teams are reorganizing around small generalist pods, why taste and judgment matter more as AI takes over execution, and how AI content affects Instagram's ranking and creator ecosystem. The conversation also covers recommendation algorithms, lessons from past product failures, handling public backlash, and parenting with screens.

01Key learnings

• Small pods of four to six generalist engineers plus a generalist 'product staff' role can move faster than large specialist teams, with senior specialists brought in only when needed.
• As building gets cheaper, spend more time deciding what to build; taste and judgment about strategy become the scarce skills, per Mosseri.
• Keep investing in tomorrow's senior talent in each function, not just today's, or the organization will regret it in a few years.
• Strategy should be opinionated and controversial enough that a reasonable person could disagree; an AI asked for strategy lazily tends to produce predictable, competitor-expected answers.
• Treat AI tools as clear-eyed: know what they do well and poorly, and develop an instinct for how their capabilities will shift in the coming months.
• Label or surface AI-generated content and account provenance rather than filtering content by the tool used; invest in individual creators because authenticity may become more valued amid synthetic content.
• Explain the trade-offs of product decisions publicly, and plan the communication strategy before running any controversial test, since tests at scale will leak.
• Recommendation systems are less semantically 'smart' than people assume; LLMs can now help translate embedding-based interest maps into human-readable descriptions that give users more agency.

02Frameworks mentioned

Summary and takeaways written by PM Atlas; quotes are short excerpts. © the original author.