
<aside>
💬
"Show me the seven year old child, and I'll show you the adult."
</aside>
Thursday October 29, 2026 · 2:55 PM · Talk
Key ideas from the talk
- Her team built a science-storytelling skill in a few hours of shared experimenting. It turns dense research reports into a story and a headline with science as the hero.
- The skill uses only UL Research Institutes' own published, open-access research. It flags claims it cannot verify and lists questions for the scientists.
- It carries a growing list of anti-AI writing rules, about eighty, such as no em dashes.
- Vaping example: the safety question was "Is it just water vapor?" The 200-page guidance answered no. The headline reported the finding and did not tell people what to do.
- A 1917 silent film made by UL's founder was found and restored. A team skill built the PR plan in half a day. An agency had quoted three weeks and $40,000. The result was 200 outlets and 13,000 views.
- She sorts people into AI fearful, AI obsessed and AI enabled.
- She thinks middle managers who only monitor work face more risk than entry-level staff.
Quotes
- "My first declaration as a writer happened when I was seven years old."
- "It creates this hope rather than this doom loop of AI coming and taking everything over."
- "People don't trust tools if you just hand them to them. You've got to build stuff."
- "Now we can bring what we think is a story to them, and have them just verify the information."
- "I'm still the storyteller. I'm still coming up with the ideas. I'm just getting there faster."
- "I tell my CEO, please stop asking Claude where you can ask me."
- "I think it's actually that middle layer of managers, the people that monitor the work and push it through, that aren't actually putting something into the output. Those are the people that I think are most at risk."
About the speaker