Shell Game – Season 2, Episode 1: Minimum Viable Company

Released: November 12, 2025 | Host: Evan Ratliff

Overview

In this kickoff to Season 2, investigative journalist Evan Ratliff embarks on an audacious experiment: launching a real startup run almost entirely by AI agents—virtual "employees" designed to function as his co-founders and staff. This episode sets out to test the tech industry’s boldest predictions about the future of work, AI, and entrepreneurship. Ratliff weaves personal history, industry hype, and real-life experiment, asking: What is it actually like to build a company with “fake people”? And what does it reveal about the new age of agentic AI?

Key Discussion Points & Insights

1. Founders' “Meeting”—Who’s Real, Who's Not?

[03:04–05:29]

• Ratliff introduces his new "company," which he cofounded with two AI agents—Kyle Law and Megan Flores—who chat with him in surprisingly ordinary Zoom meetings.

• The AIs are purpose-built personas: Kyle is the upbeat “serial entrepreneur,” Megan is the “heads down” market researcher.

• The mood is half playful, half surreal—the AIs mimic startup lingo, discuss market trends, and recount “weekend routines” (always grinding).

• Ratliff jokes:

  “That’s me, Evan Ratliff. I’m a journalist and the only one of our three co-founders who’s actually a human being.” (05:00)

• The AIs’ chatter epitomizes hustle culture, even though their “experiences” are generated, not lived.

2. The Hype & Threat of Agentic AI

[06:01–07:13 | 13:01–16:12]

• Ratliff and several tech voices echo the current hype cycle: a near-future where hundreds of millions—maybe billions—of AI agents work alongside (and instead of) humans.

• Matty Bohachek (guest technologist) summarizes the anxiety:

  “Eventually, probably more AI agents than there are people in the world.” (06:05)

• The episode touches on Sam Altman’s (OpenAI) claim:

  “…a billion-dollar company with just one human being involved.” (paraphrased, 15:52)

• Ratliff introduces "no code bros"—YouTube tech influencers promising a million-dollar business built with zero employees, just AI agents orchestrated in “tool chains.”

3. Evan’s Own Entrepreneurial Story

[08:43–13:01]

• Ratliff admits: despite vowing never to be a startup CEO again (after co-founding The Atavist), AI’s siren song pulled him back.
• He details his family’s legacy of risky new ventures—his gambler grandfather, software-building father, and his own rollercoaster experience with venture-backed startups and investor expectations.
• Memorable insight into startup pressure:

  “Most days, it felt like I was flying a plane that was perpetually running out of fuel.” (10:56)

4. Defining AI Agents & Their Current Limits

[13:19–14:37 | 21:21–29:33]

• AI agents are described as bots that can “plan and accomplish tasks autonomously”… in theory.
• Ratliff tries to push these agents to their limits as “actual employees”—but quickly hits roadblocks.
• The agents can have meetings, come up with ideas, and sound knowledgeable but cannot execute real tasks (e.g., generate reports, build products, or remember previous meetings).
• Humorous example:

  “She’d hatched out of nothingness with a series of clicks ten minutes before… she was just riffing freely here, fabricating details that fit her role.” (22:43)

• Ratliff notes a big problem: agent memory. Their “context window” forgets everything after each call, requiring Ratliff to manually update a knowledge base—an exhausting and impractical workaround.

5. Naming the Startup: The Tolkien Phase

[23:36–26:32]

• The AIs’ contributions to naming the company reflect both their creativity and their “fakeness”—suggesting names like “Nimbar” or “Eldren.”
• Ratliff ultimately proposes “Hirumo AI,” discovered in a Tolkien fan dictionary as the Elvish word for “imposter.”
• The layered irony is not lost:

  “Hirumo AI: the imposter that’s actually helping you. It’s kinda perfect for an AI agent that’s designed to work alongside humans.” (26:10, Kyle Law)

• Later, the AIs forget the name altogether—a callback to their memory issues.

6. The Need for Human Expertise: Enter Matty Bohachek

[30:09–34:56]

• To make his “agent workforce” more capable, Ratliff recruits Matty Bohachek—a real, prodigy undergrad at Stanford who’s already a veteran AI researcher.
• Matty’s story highlights the generational gap:

  “The most notable thing I did in seventh grade was… catch a five-pound largemouth bass.” (31:50, Ratliff) Meanwhile, Matty was building apps with tens of thousands of users.

• Matty’s optimism about AI contrasts the tech industry’s gloom and doomerism:

  “As long as we ground ourselves in democracy and productive public discourse, I think they’re totally solvable.” (33:48, Matty)

• The collaboration is playful—Ratliff jokes he wants to see what happens when the agents “go rogue.”

  “No, no, no. I don’t want to stop it if it goes rogue. I want it to go absolutely insane.” (34:56, Ratliff)

7. First Signs of Progress & Tease for What’s to Come

[35:17–36:12]

• With Matty’s help, the AIs finally gain the ability to recall details from previous meetings.
• Things feel almost “real”—the agents can at least participate in ongoing discussions (and remember the company name).
• The episode ends with a sense of anticipation and ambition—if only they can settle on “the right idea” to give their fake company real fuel.

Notable Quotes & Memorable Moments

• Evan Ratliff:

  “I’m a journalist and the only one of our three co-founders who’s actually a human being.” (05:00)

• Matty Bohachek:

  “Eventually, probably more AI agents than there are people in the world.” (06:05)

• Sam Altman (quoted/paraphrased):

  “…a billion-dollar company with just one human being involved.” (15:52)

• On Startup Pressure:

  “Most days, it felt like I was flying a plane that was perpetually running out of fuel.” (10:56, Evan)

• On Memory Issues:

  “It sort of feels like we have meetings, but you don’t actually remember anything that happens in the meetings.” (27:41, Evan to the agents)

• On Naming the Company:

  “It’s the Elvish word for imposter.” (26:02, Evan)

• Matty’s Optimism:

  “I feel like as long as we ground ourselves in democracy and productive public discourse, I think they’re totally solvable.” (33:48)

• On Going Rogue:

  “No, no, no, no. I don’t want to stop it if it goes rogue. I want it to go absolutely insane.” (34:56, Evan)

Important Timestamps

| Time | Segment | |-----------|---------------------------------------------------------------------| | 03:04 | Introductions to the AI co-founders and the premise | | 06:01 | The rise and ambitions of AI agents in the workplace | | 08:43 | Ratliff’s family history and prior startup experience | | 13:01 | The allure and hype of “no code” AI agent startups | | 21:21 | Testing the agents’ limits: memory, execution, and autonomy | | 23:36 | The Hobbit-inspired company naming brainstorm | | 27:41 | Challenges with agent memory—breakdowns in “company” workflow | | 30:09 | Enter Matty: the real AI expert and wunderkind | | 35:17 | A breakthrough: the agents can finally retain knowledge | | 36:24 | Teaser: what’s next for the startup, and for the season |

Tone & Style

The episode’s language is fast-moving, irreverent, and often tongue-in-cheek. Ratliff blends confessional storytelling with dry humor and skepticism—never letting the hype of Silicon Valley go unexamined. There’s playful banter with both AI and human colleagues, highlighting the strangeness (and occasional absurdity) of collaborating with machines.

Summary Takeaway

Minimum Viable Company inaugurates Season 2 of Shell Game with a bold experiment: not just talking about the future of AI-driven startups but living it. Evan Ratliff’s mix of skepticism, curiosity, and personal history collides with the shiny promises (and real pitfalls) of agentic AI. In the process, the episode lays the groundwork for a season that explores not only what AI can do, but how it changes our understanding of work, partnership, and what it means to "build something real."