Issue No. 46Week ending Sunday, November 15, 2026844 episodes · 3,675 write-ups
The Podcast Summary.

The best ideas from 30+ top podcasts, every week.

AI agents

Mona Sarantakos on AI agents

6 quotes from 1 episode on Lenny's Podcast, each with a timestamped link to the source.

6 quotes1 episode

The short version

Mona Sarantakos stated that personal AI agents succeed when users feel complete ownership. To debug continuous back-and-forth interactions, engineering teams deploy autonomous agent swarms to confirm product bugs before assigning developers.

Most interesting insights

Internal teams repeatedly pushed back against launching a personal agent with a stuffed animal avatar.

“Jolly is like our cockroach. It just like never dies. Every few months, somebody would be like, there's no way you're going to launch this product with a stuffed animal at the top.”

Mona Sarantakos, Lenny's Podcast · November 2026 · Watch at 0:47 ↗

From Why Meta Chose a Stuffed Animal Over Its Corporate Logo

A personal agent functions as an asset owned directly by the user.

“The agent needs to work for you. You are the owner.”

Mona Sarantakos, Lenny's Podcast · November 2026 · Watch at 29:12 ↗

From Why Meta Chose a Stuffed Animal Over Its Corporate Logo

Top talking points

  1. Agent swarms identify product bugs in multi-turn conversations

    Multi-turn conversations created novel back-and-forth patterns without existing training data. Mona Sarantakos explained that engineers deployed automated agent swarms to verify genuine bugs before assigning human developers.

    “And then when we started building this, it was like completely multi-turn, right? Like it was like back and forth, back and forth, back and forth. And we don't have data that looks like that.”

    Mona Sarantakos, Lenny's Podcast · November 2026 · Watch at 39:46 ↗

    From Meta Uses Agent Swarms to Debug Multi-Turn AI

    “The engineers were like oh we'll swarm it and figure it out. Swarmed it, and like, oh no, this is an actual problem that we need to solve, and we're going to get somebody on it.”

    Mona Sarantakos, Lenny's Podcast · November 2026 · Watch at 40:46 ↗

    From Meta Uses Agent Swarms to Debug Multi-Turn AI

  2. Personal agents require distance from corporate branding

    Users avoid talking to corporate logos during personal interactions. Mona Sarantakos stated that Meta designed the Muse agent so users feel complete personal ownership.

    “But I think when you're talking to something that feels so personal, the last thing you want to talk to is a corporate logo.”

    Mona Sarantakos, Lenny's Podcast · November 2026 · Watch at 0:56 ↗

    From Why Meta Chose a Stuffed Animal Over Its Corporate Logo

    “And this really kind of gets into this concept that we have about like the Muse is yours and like we wanted to really be yours and not like a meta entity necessarily.”

    Mona Sarantakos, Lenny's Podcast · November 2026 · Watch at 25:46 ↗

    From Why Meta Chose a Stuffed Animal Over Its Corporate Logo

Key takeaways from these write-ups

Why Meta Chose a Stuffed Animal Over Its Corporate Logo

  • Lead designer Alex Cornell built the original prototype of Meta's personal agent, Muse, with a one-shot prompt that produced a stuffed-animal creature named Jolly.
  • Mona Sarantakos nicknamed Jolly the team's "cockroach" because executives spent months insisting the team replace the character with standard corporate branding.

Meta Uses Agent Swarms to Debug Multi-Turn AI

  • Static coding benchmarks fail on consumer agents because real user behavior spans multi-turn, multi-day interactions rather than single prompts.
  • Mona Sarantakos deployed autonomous agent swarms at Meta Superintelligence Labs to inspect live user conversations and generate synthetic edge-case data.

How we attribute quotes. Every quote was matched against the episode transcript, so the words and the timestamp are real (we trim filler words like "um", nothing else). The name comes from our written summary of the episode, and we use it only when a separate check of the captions finds that person on the episode. YouTube gives us no voice-by-voice transcript, so open the timestamp to hear who is talking. See a wrong name? Tell us and we fix or remove it.

More

The Sunday Email

Get next Sunday's issue in your inbox.

The week's best AI, tech, business, science, and culture podcasts in one short read. Free, every Sunday.

You get both newsletters, The Throughline and The Carry. No spam. Unsubscribe with one click.