Adam Mosseri: AI is a tailwind for authenticity
TL;DR
AI content is a tailwind for authenticity: In a world flooded with synthetic content, people will seek out real creators more, not less.
Teams are shrinking from 12-15 to 6 people: Smaller pods with generalist 'product staff' move faster with less coordination overhead.
Taste matters more as building gets easier: The premium shifts from execution to deciding what to build; designers with taste are uniquely valuable.
The algorithm isn't as smart as people think: It's embedding correlations, not semantic understanding, and LLMs are only now making those readable.
Chronological feeds create bad incentives: Pure chronological would drown feeds in institutional spam; algorithmic drives higher satisfaction despite user complaints.
Curators beat visionaries for product leadership: The best leaders create environments where great ideas bubble up rather than trying to generate everything themselves.
The Breakdown
From Baker's Dozen Teams to Six-Person Pods
Instagram is restructuring from teams of 12-15 specialists (Android engineers, iOS engineers, server engineers, a PM, designer, data scientist) down to pods of 4-6 generalist engineers plus one 'product staff' role. The smaller teams move faster with less coordination overhead and fewer design-by-committee decisions.
Product Staff: The Generalist Who Can Do Everything
The new 'product staff' role is a hybrid that blends PM, design, data science, and research responsibilities. Mosseri notes that AI tools now let generalists do work that previously required specialists, like pulling waterfall analyses. Senior designers and data scientists are converting to product staff to expand their influence across functional boundaries.
Taste Over Execution in an AI World
As AI makes building easier, the premium shifts from execution to deciding what to build. Mosseri is 'long on designers' because they tend to have taste, which he sees as harder to automate than technical skills. He notes that AI-generated code has a recognizable 'vibe' that makes taste more valuable, not less.
What the Algorithm Actually Knows About You
Mosseri clarifies that people overestimate how much the algorithm semantically understands their interests. For years it was just embedding-based correlations, giant vectors in dimensional space that happen to correlate with surfing or coffee. Only now are LLMs making these illegible artifacts readable, letting users 'see their algorithm' and adjust topics.
Why Synthetic Content Drives People Toward Authenticity
Mosseri argues that AI content will be a tailwind for Instagram because in a world flooded with synthetic content, people will seek out authenticity and real creators more, not less. Instagram's focus on the person behind the content positions it well. He doesn't want to filter AI content but believes users should know if content is AI-generated and who posted it.
Catching Up to TikTok's Creator-Breaking Power
Mosseri credits TikTok for inspiring Instagram's investment in exploration-based ranking, which tests content to find new interests rather than just exploiting known preferences. This approach helps small creators break out. Instagram has been playing catch-up but now has line of sight to best-in-class recommendations for the first time in his tenure.
Being the Face of Controversy and Surviving the Comments
Mosseri actively engages critics on social media because he believes the debate happens with or without Instagram's participation. He recalls being a 25-year-old designer devastated by the first homophobic, anti-semitic comment on his News Feed redesign. His coping strategy: perspective. If you spent 50 minutes daily at your desk and he rearranged it without warning, you'd be pissed too.
Screen Time Boundaries with an AI-Literacy Twist
Mosseri's kids earn iPad time through homework, with approved apps and weekend windows. But he's also teaching his 10-year-old to 'vibe code' using Claude, building a 19-level platformer game together. He wants kids digitally and AI literate, not sheltered, because falling behind on these skills is the real risk.
Was This Useful?
Share
Keep Reading
Make Alcreon Yours
Tune your feedFive quick questions, and the feed ranks what matters to you first.Or just get notified
The weekly Echo. Signal worth keeping in your inbox.
Every new piece, announced on X.
Read Next
See all
Playbook
The Retirement Email Isn't a Warning
Model retirements now arrive every few weeks; the config-eval-rehearsal loop turns each deprecation email from a fire drill into an afternoon swap.
Playbook
The Cheapest Model That Passes
OpenRouter lists 400 models behind one API. The fix for choosing isn't a better leaderboard, it's a four-step protocol that ends in a real eval.
Playbook
Cheap Models, Hard Tasks
Most agent workflows route every step to the frontier model by default. The bill scales with how chatty the agent gets, even when most steps don't need that brain.