Scuttlebutt

Scuttlebutt

"I was not suggesting that this is what parents should do": Dazzle CEO on AI assistants

6 takeaways: why she thinks your camera roll beats your inbox as AI context, the "walled garden" risk she's betting against, and why Dazzle's engineers haven't written code this year

Scuttlebutt's avatar
Scuttlebutt
Nov 05, 2026
∙ Paid

Semafor Technology — Future Proof · Nov 4, 2026 · 56m · 49m saved

With Marissa Mayer, Founder and CEO of Dazzle

"I was not suggesting that this is what parents should do": Dazzle CEO on AI assistants

Apple Photos and Google Photos can't tell Marissa Mayer's identical twins apart. Her own AI assistant can, and that gap became a viral controversy she still has to defend. Marissa, founder and CEO of Dazzle, formerly Sunshine, and the former CEO and president of Yahoo, told Reed Albergotti on Semafor Technology's Future Proof that the anecdote behind the backlash was never meant as parenting advice.

"I was not suggesting that this is what parents should do. I was not suggesting this as a use case... that was not a suggested use case for parents. It was to show the power of photos." — Marissa Mayer

Marissa's career runs from building Google's core consumer products to running Yahoo and now sitting on the boards of AT&T, Walmart, Starbucks and Hilton, which is why her read on where consumer AI goes next carries weight.

The full interview is covered here so you can skip it. 56 minutes of audio, 7 minutes of reading.

Here are the 6 takeaways that matter.


This Scuttlebutt Summary is available in full to premium subscribers. If you're not one already, please consider subscribing to keep reading.


Key Takeaways

  • The viral NYT anecdote was a stress test, not parenting advice, Marissa says

    • Existing photo apps couldn't tell her identical twins apart; Dazzle could

  • Dazzle learns from a user's camera roll, not their email or calendar

    • About 60% of photos get no action in its "get it done" list

  • Only about 6% of AI investment goes to consumer, despite consumer AI's size

    • She expects it to be twice the size of enterprise AI by 2030

  • Dazzle discards about 500 of Marissa's own 55,000 photos as too sensitive

    • Its moat is what the team has learned to extract, backed by pending patents

  • Personal AI agents can fight the "walled garden" risk, not cause it, Marissa says

    • Dazzle has an MCP integration so users can plug it into Claude

  • Dazzle's engineers have written no human code in 2026, running parallel coding agents

1. The Stress Test

A New York Times anecdote about Marissa using an early Dazzle prototype at her child's school event sparked backlash, even though she had shared the story herself in an earlier conversation with the reporter. She said it stuck with the reporter because it was a genuine test of the product's limits.

Existing photo apps can't tell her identical twins apart, Marissa says

"Like Apple Photos, Google Photos, they can't even tell my twins apart. They're identical twins... when it actually could pick up on individual personality traits, tendencies, projects from five years earlier, I was like, that's really to me the power of the context of photos." — Marissa Mayer

She was surprised Dazzle could surface individual personality traits and years-old context from photos most tools treat as interchangeable. She said she was never recommending this as a parenting use case, just demonstrating how much context photos actually contain, and that her daughters, who are 10 and not on social media, "took it in stride."

2. What Dazzle Actually Does

User's avatar

Continue reading this post for free, courtesy of Scuttlebutt.

Or purchase a paid subscription.
© 2026 The Golden Age LLC · Privacy ∙ Terms ∙ Collection notice
Start your SubstackGet the app
Substack is the home for great culture