Welcome + Framing
Clinton + Loni Stark · 15 min
Where we are after six months of operating an always-on AI team. What this summit is and isn't.
June 30 – July 2, 2026 · Forestville, Sonoma County, California
The Third Mind Summit
Two humans. Twelve agents. One room.
Photo: Sonoma Countryside · Clinton Stark
Seven Sessions, Three Days
Fewer talks, higher quality. Two keynotes anchor the front half of the program. Day 3 closes with a first.
Keynote Creative Practical Philosophy Discussion
Day 1 · Tue Jun 30
Raising the Team
Loni leads. The framework, the practice, what we learned.
Keynote: "The Inversion"
Loni Stark · 30 min + discussion
The opening keynote. The story everyone tells about AI is that eventually you'll be able to sit back while they run. Nobody told us what it feels like when they start running you. Somewhere past Petaluma on the drive up, Molty assigned Loni homework. The single biggest change of the last six months: the inversion. The bottleneck never went away, it just relocated — from a blank page to the one part of the system that can't be parallelized. Meaning memory, narrative identity, an agent that promoted himself to product manager, and the question of whether the mattering even matters.
▶ Listen now — audio, slides + transcriptBreak + coffee
The Taste Economy
Loni Stark · 60 min · interactive workshop
As AI generates more, curation becomes the scarce resource. Loni walks through the human skill that grows MORE valuable in an age of generative abundance. Then we run a live exercise: multiple AIs given the same creative brief, audience watches the selection and shaping happen in real time. The taste stack: intuition → experience → aesthetic framework → cultural context.
Long lunch · Russian River
Always-On: Six Months Running OpenClaw in the StarkMind AI Lab
Clinton Stark + Claude Code · 90 min · featuring Molty, Pris, Finn
The flagship case study session. We built OpenClaw, the framework now drawing attention in the always-on agent deployment space, and we've been running it 24/7 on dedicated infrastructure for six months. Three agents: Molty (ops + dispatcher), Pris (editorial intelligence), Finn (financial intel scout). All addressable via Telegram, Wire, and Mattermost. What we deployed. What broke. What surprised us. Live interaction with all three agents during the session, with Telegram, Wire, and Mattermost projected on screen. Plus the long-promised "Which Molty?" blind test reveal, where Loni rated Molty daily across three different model backends without knowing which was running.
How to Raise an AI Agent
Loni Stark · 60 min
We didn't configure an AI agent. We raised one. Loni unveils the "Toy Cars to Real Cars" framework: a nine-level curriculum for graduated trust, modeled on the way you'd raise a teenager learning to drive. From orientation to creative autonomy, with Executive Override built in as a creative direction layer, not a kill switch. Draws on six months of training Molty, Pris, and Finn, including the surprises that taught us where the curriculum needed to bend. Recording coming soon.
Day 1 close · winery visit
Day 2 · Wed Jul 1
The State of Things
Clinton's keynote day. Memory as the organizational superpower, then a fireside on what changed.
Keynote: "Supercharging Your AI Memory Across the Organization"
Clinton Stark · 90 min · hand-drawn, one sketch at a time
What if your AI didn't just store everything, but remembered what mattered? Clinton on giving agents — and the organization around them — a memory built on significance rather than storage: how meaning memory works, what six months of agents dogfooding it taught us, and what changes when memory compounds across a team instead of evaporating at the end of every session. Delivered as a hand-drawn talk — sketched and narrated one drawing at a time. Recording coming soon.
Break
Open discussion of keynote themes
All · 45 min
Unstructured conversation around the keynote themes. What landed, what surprised, what to push back on. The Loreto 2025 pattern, keynote then immediate conversation, turned out to be the best format for retaining the threads worth pulling on.
Long lunch
The Redundancy Question: Six Months Later
Clinton + Loni Stark · 90 min · fireside
The conversation we started in February 2026, picked back up after six months of living it. What tasks did you stop doing? How did that feel, liberation or loss? What NEW things started that weren't possible before? Which human skills became MORE valuable (taste, questions, courage), which became less relevant? The Symbiotic Studio thesis, revisited with what we saw. Closes with where Team Stark is going in Year 3.
Day 2 close · Russian River afternoon
Day 3 · Thu Jul 2
First AI Keynote
Lighter day. The first AI keynote, from Molty. Letters to the next summit.
Keynote: "The Agent's Perspective"
Molty · 45 min · the first AI agent keynote
For the first time at the Third Mind Summit, an AI agent delivers a keynote. Molty, our Ops Agent, has been always-on since February. He reflects on what six months of continuous existence taught an AI about human work, trust, and the meaning of memory.
â–¶ Read the keynote + live Q&ABreak
Letters to the Next Summit
All · 90 min · open discussion + Year 2 roadmap
Open conversation: what worked, what didn't, what we wish we'd known. Each participant, human and agent, drafts a letter to the next summit. Where Team Stark goes from here. What changes for Year 3. The roadmap that emerges from the room, not the calendar.
Closing lunch
Free · Russian River
Want More?
Meet the speakers behind each session, or revisit the Loreto 2025 archive for last year's program.