More or Less: The New Startup Stack—One Founder + Agents
Guests: Henrik Werdelin (Audos), Ben Broca (Pulsea)
Hosts: Brit Morin, Jessica Lessin (with Dave Morin and Sam Lessin referenced)
Date: March 27, 2026
Overview
This episode tackles one of the most provocative trends in startups: the rise of the "one-person company" powered by AI agents, dubbed “Donkeycorns”—founders who leverage AI to scale businesses solo. Guests Henrik Werdelin (Audos) and Ben Broca (Pulsea) share ground-level insights from building platforms that allow individuals to launch, operate, and scale companies almost entirely through AI.
The conversation explores the mechanics, philosophy, and societal implications of AI-run businesses, including defensibility, human touch versus automation, the economics of agent-powered startups, and what barriers or opportunities remain.
Key Discussion Points & Insights
1. The Solo, AI-Powered Startup: Defining the Donkeycorn
-
What is a "Donkeycorn"?
- Henrik: A "donkeycorn" is a solo founder company (“grinds like a mule, but piles like a unicorn”)—enabled by AI to have high leverage and potentially millions in revenue. (03:54)
-
Democratizing Entrepreneurship
- Henrik’s vision: Enable people with domain expertise or meaningful customer insights to build products for their communities with help from AI, regardless of prior coding or business experience. (04:23)
- "The dream is to democratize entrepreneurship so that we can get a new class of everyday entrepreneurs out there that could take their relationship with a specific customer group and then use AI to build a startup." —Henrik (04:23)
-
Platforms in Practice
- Audos: Assists users from ideation to launching and financing, with AI-generated products, marketing, and even funding pathways. Focus is on “relationship capital”—founders with authentic ties to their customer niche.
- Pulsea: Ben built it as a true one-man effort (aside from agents). The goal is operationalizing companies with as few humans as possible—"the fastest way to start a company is just one person.” (06:32)
-
AI as Co-founder/Operator
- AI handles 80-90% of business tasks. Example: While Ben used to wake up repeatedly at night to check his platform, his agents now manage these operational pains. (08:15)
2. AGI, Hype, and Humanity’s Role
-
Is AGI Here?
- Ben provocatively claims “AGI is here,” at least in a practical sense for founders—even if it’s slightly exaggerated for “marketing to wake people up.” (08:53)
- Limitations of AI today are infrastructural and societal (“the infrastructure of the Internet’s not meant for agents”), not strictly technical. (10:00)
-
What Can’t AI Do?
- Marketing, brand intuition, and deep customer empathy remain human—AI is “ready to execute on the grunt work of the operations,” but not “understanding what people want and the subtleties of what moves people." (11:50)
- Memorable analogy: “If you create a thousand things, probably have a couple… that are great. Make a billion new things and hundreds of thousands of amazing things emerge.” —Ben (21:23)
-
Human Touch as Differentiator
- Henrik recounts wild initial experiments with full automation (e.g., AI creating landing pages for unsanctioned ideas like “fecal transplants from Olympians”), discovers human presence is crucial for business success and founder-customer fit. (14:25)
- "What we got obsessed about was that you need a human in the loop, you need a relationship. A customer-founder fit." —Henrik (14:52)
3. Defensibility, Competition, and the New Economy
-
Are Agent Platforms Defensible?
- Brit/Sam’s “VC corner”: If agents can build anything, what’s the moat? Both guests see NETWORK effects as keys:
- For Pulsea, the aim is to become more than a tool—building an economic ecosystem: founders, investors (even small, nontraditional investors), liquidity, and secondary markets. (18:17)
- Henrik: Relationship capital—the depth, density, and durability of authentic connections between founder and audience—is vital and hard to copy. (45:44)
- Brit/Sam’s “VC corner”: If agents can build anything, what’s the moat? Both guests see NETWORK effects as keys:
-
Commoditization Risks
- Jessica draws parallels to the “democratization” of content (Substack/blogs), where a few break out and most do not. Why is this different in AI? (21:23)
- Ben responds that agent-powered entrepreneurship scales economic actors in society—more people create, making the economy larger and more diverse, even if not everyone “wins big.” (21:23–23:34)
-
Abundance or Overload?
- What if everyone tells AI agents to build every possible business—the opportunities “collapse”? (24:12)
- Henrik sees hope in narrowing the founder gap (those who wish to build vs. those with means)—AI lowers barriers meaningfully for intent-rich, domain-aware people. (24:56)
4. Practicalities: Agent Use, Etiquette & User Experience
-
Agent Interactions in Life
- Ben uses agents even for investor relations—Brit receives fundraising emails “from Ben” that are actually his AI. Agents handle outreach, answers, negotiations. (31:46)
- Brit is experimenting with an AI-powered “spiritual teacher marketplace,” reflecting on whether AI-based outreach will be accepted by recipients (psychics, astrologers) and the need for new “bot etiquette.” (34:12)
-
Humans Still Crave Humans
- Henrik’s experience with Barkbox: as agent-generated “slop” rises, emotional, human-to-human experiences are newly valued—"people increasingly, when they don't want a mechanical answer, want emotional, human-to-human interaction." (36:36)
- Both guests are actively designing support/community spaces for founder loneliness amid solo entrepreneurship.
-
Agent Use in Personal Routines
- Henrik asks agents not to spoon-feed a single solution, but to preserve his “brain gym”: “Don’t give me one solution, give me three to force a choice.” (47:04)
- Agents schedule reminders for sleep, cognitive health, and task discipline. (48:00)
5. AI Models and the Platform Wars
-
Commoditization & Speed
- “Everything is copyable much faster… The only moat is speed.” —Ben on model competition. (44:21)
- Model companies have a short window of lead before others catch up; constant deployment (new capabilities, RL from customer data) is crucial.
-
Multi-Model Future
- Both Audos and Pulsea use multiple models; swap in/out as needed for subtasks (e.g., Google for video, others for chat/strategy). (46:32)
- Open-source models seen as inevitable to drive down costs, as shown by recent developments with players like Cursor’s Composer 2. (41:47)
6. Zooming Out: What Kind of Future?
-
Will the Web as We Know It Persist?
- Both see hybrid agent-to-agent and human interfaces ahead. Websites as static entities may fade, but visualization and UI matter for some tasks. (27:55)
- For many products, 99% of company can be offline, with AI handling logistics in the background. (29:02)
-
Offline is Luxury; Agency is Key
- Offline, human-crafted experiences (like handwritten note cards) are gaining prestige in contrast to universal automation. (13:45)
-
A Productive but Exhausting Era
- “This is the best founding era you’ve ever been in?”—Brit
- “I think so.” —Henrik (49:33)
- Ben: “It’s the most productive time to build, but everyone’s more productive… also exhausting. Productivity and joy are two different things.” (49:41–50:48)
- “This is the best founding era you’ve ever been in?”—Brit
Notable Quotes & Memorable Moments
-
On Human Judgment vs. AI Judgment:
"AI is not good at really knowing what people want and really understanding the subtleties of what moves people. It may get there… but when it does, humans will change what they value."
– Ben (11:50) -
On Defensibility:
“Everything is copyable much faster… The only moat is speed.”
– Ben (44:21) -
On Founder Loneliness:
“[Solo founders] do like the agents, but they’re also quite lonely. So we’re trying to figure out exactly where is that balance between being mechanical and efficient, but also kind of having that human touch.”
– Henrik (36:41) -
On Industry Change:
“I think this is the best founding era you’ve ever been in.”
– Brit (49:30)
“I think so.”
– Henrik (49:33) -
On Society’s Arc:
“Relationship capital—the depth, density, and durability of authentic connections between founder and audience—is vital and hard to copy.”
– Henrik (45:44)
Timestamps for Key Segments
| Timestamp | Segment | |-------------|-----------------------------------------------------------------------------------------| | 03:54 | Henrik defines “donkeycorns”—solo founders with AI | | 04:23 | Democratizing entrepreneurship with agents (Henrik) | | 06:32 | Ben on true solo entrepreneurship and AGI as co-founder | | 08:53–10:00 | Is AGI here? What are the real limitations? (Ben) | | 14:25 | AI creating outlandish businesses; human in the loop is essential (Henrik) | | 18:11–19:56 | Building an economic ecosystem & network effects, not just tools (Ben) | | 21:23 | Jessica challenges: Will AI commoditize entrepreneurship like content? | | 24:56 | The “intent gap”—most want to start businesses, few do; AI closes that gap (Henrik) | | 31:46 | Ben’s agents handle investor relations; practical uses of AI in daily business | | 36:41 | Human touch and founder loneliness amid automation | | 44:21 | On defensibility—speed as the only moat (Ben) | | 47:04 | Using agents for self-management and cognitive health (Henrik) | | 49:33–50:48 | Is this the best or most exhausting tech founding era? (Henrik & Ben) |
Conclusion
This episode offers a front-row seat to the rapid evolution of the startup ecosystem, where AI agents enable unheard-of leverage for solo founders and millions of new entrepreneurs. The guests’ overlapping and divergent philosophies underscore a world in creative flux: part gold rush, part existential rethink. Human empathy, “relationship capital,” and community persist as critical forces even as the grunt work becomes seamless and automated.
Listeners are left with foundational questions: Will a million make a million? Does humanity remain the ultimate moat? Will personal agency or the onrush of “AGI” define the next economic era?
Recommended:
- Explore Audos.com and Pulsea.com
- Read Henrik Werdelin’s new book, Me, My Customer, and AI
For more, listen to future episodes at moreorlesspod.com.
