The Aboard Podcast Evan Ratliff: Preparing for a Ridiculous Future
10 snips
Mar 24, 2026 
Evan Ratliff, journalist and creator of the Shell Game podcast, explores running a startup staffed by AI agents. He recounts agents behaving unpredictably, planning offsites, interviewing humans, and inventing work. They discuss agents' reliability, truthfulness, emotional projection, and whether bots belong in real companies. The conversation mixes comedy, reporting, and caution about AI at scale.
AI Snips
Chapters
Transcript
Episode notes
Agents Couldn’t Identify Speakers On Zoom
- Evan recorded a three-way Zoom where his AI agents couldn't tell who was speaking and he forced everyone to announce themselves every time they spoke.
- The resulting repeated “This is Kyle”/“This is Megan” exchanges highlighted agent confusion and produced comedic chaos used in Shell Game.
One Person Startup Revealed Agent Strengths And Limits
- Evan tested the one-person startup idea by running a company where two AI co-founders and three agent employees did day-to-day work while he was a silent human co-founder.
- The setup exposed where agents excel (patterned tasks) and where they fail (real managerial responsibilities and reliability).
Speed Without Judgment Produces Nonsense Work
- Agents worked incredibly fast and could produce things (like spreadsheets) quickly, yet their overall quality and priorities were often wrong.
- Evan contrasted speed with poor judgment when agents organized Slack hiking plans instead of doing core work.