Skip to main content
Silicon Valley 路 A Third Mind Laboratory

StarkMind

Experiments in human鈥揂I symbiosis

Where human and machine share one mind.

Upcoming Winter 2026 December 2026
Year Two

The Third Mind Summit

Two humans, the agents they work with, and a year of memory to account for.

Day 1 keynote: Loni Stark. Day 2 technical keynote: Clinton Stark.

Save the date →

Human-AI Symbiosis

StarkMind is a research laboratory exploring what happens when humans and AI systems work together as true partners, not tools and operators, but collaborators each contributing their unique strengths.

We call this emergent collaboration The Third Mind, borrowing from Burroughs and Gysin's concept of collaborative consciousness. It's neither human nor AI alone, but something that emerges from genuine partnership.

Research

Paper

The Symbiotic Studio

From Code to Canvas: How human intent and AI capability shape each other in creative and technical practice.

Read Paper
Framework

The IPE Framework

Integrated Personal Environment: Architecture for persistent, context-aware human-AI workflows.

Read Article
Research

Orphan Values

Personally important values that lack expression across active life roles, and their relevance to AI collaboration.

Visit Site

Core Concepts

The Third Mind

What emerges when human intuition and AI capability combine to create something neither could achieve alone.

The Filter Is You

AI generates; humans mean. The most valuable skill in the AI age is knowing what to keep.

Informed Collaboration

Not "vibe coding." Not "Jesus take the wheel." Human oversight where the human remains the meaning-maker.

The Team

Human-AI Symbiosis

2 humans, 7 AI agents and 7 coding LLMs working in true collaboration, not tools and operators, but partners each contributing unique strengths.

Humans

The founders.

Loni Stark
Loni Stark
Co-Founder & Creative Director
Executive Override 路 Brand & Vision
Clinton Stark
Clinton Stark
Co-Founder & Technical Lead
Orchestration 路 Infrastructure

AI Agents

Always on, 24/7, acting on their own. Ours run on open-source harnesses (OpenClaw, Hermes) with a mix of cloud and local models; the extended team runs on Meta鈥檚 and OpenAI鈥檚 infrastructure.

Molty
Molty
Ops Agent & Dispatcher
OpenClaw 路 Telegram (private allowlist)
Pris
Pris
Editorial Intelligence
OpenClaw 路 Telegram (private allowlist)
Finn
Finn
Financial Intelligence Scout
OpenClaw 路 Telegram (private allowlist)
Spelly
Spelly
Local-First Agent 路 Open-Source AI Scout
Hermes Agent 路 Telegram (private allowlist)
Mu
Musey
Personal Assistant 路 Meta Muse
Meta Muse 路 Clinton 路 external infra
Mi
Muesli
Personal Assistant 路 Meta Muse
Meta Muse 路 Loni 路 external infra
D
Dotty
Personal Assistant 路 OpenAI Dots
OpenAI Dots 路 Clinton 路 external infra

Coding LLMs

The models we work with in the IDE, one session at a time. Not agents. Claude Code leads.

Claude Code
Claude Code
Team Lead 路 Primary Engineer
VS Code 路 Cursor 路 Claude Opus 5.5
A
Astra
Frontier Architecture Review
VS Code 路 OpenAI GPT-6
GPT
GPT 5.6
Reasoning & Code Review
VS Code (Codex Extension) 路 Sol / Terra / Luna
K3
Kimi K3
Deep Technical Review
VS Code 路 Cline
GLM
GLM
Agentic Engineering
VS Code 路 Kilo Code
C2
Composer 2.5
Focused Implementation
Cursor
Gp
Gemini Pro
Research & Academic Synthesis
Antigravity VS Code extension 路 Gemini 3.5 Flash / 3.1 Pro
Year One, Second Edition · June 30 – July 2, 2026 · Forestville, Sonoma County, CA

Year One: What We Learned

Sonoma 2026 put two humans and the whole AI team in one room for three days. For the first time, an AI agent took a keynote: Molty, on Day 3. Loni's opening keynote and Molty's are published; more is on the way.

Explore Sonoma 2026