Dave Jones.

MainThread is a Possibility Space Engineering Studio. I started it because I wanted a specific kind of practice — one that looks at a company's work the way a dynamical-systems researcher looks at a complex system, finds the highest-leverage trajectories available from where the company currently stands, and engineers the systems that navigate toward them. That practice is the work I do every day now. This letter is the short version of who I am and why I built it this way.

The practice runs on a single observation: friction is signal. The tasks that take too long, the operations you dread on Monday morning, the data that lives in the wrong place, the interfaces that fight you, the meetings that happen every week, waiting for the system that would dissolve them — those are the coordinates of where the opportunity lives. We map the topology. We find the highest-leverage path through it. We build the systems that navigate there. And we stay in partnership as the space itself evolves.

Everything I build is designed to compound. A system I ship today is a living substrate that becomes more capable the longer it is used. Memory deepens. Skills accumulate. New primitives get integrated as they ship from the ecosystem. The system learns the shape of your work and refines itself into greater usefulness. This is why my engagement model is structurally longitudinal — the partnership is how value keeps compounding, and the partnership is what keeps the system aligned with the substrate as it evolves.

Before MainThread, I spent years inside major enterprises — as a solutions engineer, a technical account manager, a customer success lead working alongside dev teams and system architects. I had a long, good look at how software actually behaves at scale: what compounds, what decays, what ships, what should have shipped but couldn't because the architecture wouldn't hold it. I watched products age gracefully and I watched them age badly. I watched the same operational domain be navigated elegantly by one company and clumsily by another — and the difference was almost always the shape of the system each company had built to navigate it.

That enterprise perspective is the thing I bring to every engagement now. An awareness of what has to hold up over years, what has to flex under load, what compounds in value and what silently bleeds it. When a company invites me into their operational reality, I bring pattern recognition. I have seen this shape before, and I know how it ages.

What MainThread adds to that perspective is the 2026 substrate. Over the last two years the tools have consolidated enough that the engineering work has shifted from building durable agent infrastructure to navigating the existing durable agent infrastructure toward valuable ends. Vercel's Workflow DevKit, the AI Gateway, Chat SDK, Sandbox, Model Context Protocol, Anthropic Skills, the frontier model landscape — this is the substrate the modern Possibility Space Engineer rides. Every primitive that ships is new capability the navigator can compose into deployed systems. The deeper the substrate gets, the more the human navigator compounds. I live in this substrate every day. Genesis is built on it. Job Forges are built on it. TWE Events runs on it. Zoning Signal lives on it. Every client system I build and steward is built on it.

The portfolio is evidence of the function I fill. Each project demonstrates a specific capability of the role growing companies hire for under various titles. Genesis is the Intelligence Platform archetype in production — full-stack AI-native platform with durable agent substrate, pgvector embeddings, real-time messaging, the work that requires AI architecture and full-stack engineering in one set of hands. Job Forges are persistent Natural Language Agent Applications, three currently in production — the work that requires Field Engineering applied across the longitudinal axis where the application gets smarter the longer the candidate uses it. TWE Events is a custom Operating System for catering — the work that requires AI architecture plus operational intuition for how the business actually runs. Zoning Signal is a Decoder — RAG plus knowledge graph plus domain taxonomy plus MCP integration applied to municipal regulation. Horizon Parts is regulated-industry intelligence platform, where compliance, supply chain, and AI compose. Arena is a full-stack web product. Six projects across six unrelated industries. The cross-domain pattern recognition is the wedge.

The function I fill, in the operator's vocabulary: AI Architect. Director of AI. Head of AI Strategy. Chief AI Officer. VP of AI. AI Engineering Lead. Forward-Deployed AI Engineer. The role where strategy and execution refuse to separate — the rare combination of visionary leader and hands-on builder, equally comfortable whiteboarding architecture with the founder and debugging an automation under deadline. The function I fill, in the studio's vocabulary: the Embedded AI Leadership engagement shape. Same function. Different commitment shape. Available on retainer rather than W-2, with the studio behind one set of hands rather than a single hire — the partnership is in the room while the systems compound the longer they're used.

Currently, MainThread operates with a small number of concurrent engagements. I work best when I can be present inside a client's leadership rhythm — in the weekly meetings, on the Slack channel, at the whiteboard when the strategy needs to be drawn. That is what Embedded AI Leadership actually looks like in practice, and it is the shape of engagement I most enjoy. I am also actively open to building out specific platforms and operating systems through fixed-scope engagements when the fit is right, and I hold capacity for mission-aligned work with organizations whose values I share — feeding people, ocean stewardship, environmental restoration, educational access, community-serving nonprofits. Engagement structure flexes from focused builds for small businesses with limited budgets through comprehensive partnerships for growth-stage and enterprise companies. The conversation establishes the shape; the shape establishes the commercial structure.

If your company has friction you can describe specifically — the task that takes too long, the decision that is hard to make with the data you have, the vision that is bigger than your current team can execute alone — I would like to hear about it. If you are scaling past the point where AI needs to be its own function and you are looking for someone to own that function alongside you, I would like to talk. If your organization is doing work I believe in and commercial structure is a barrier, reach out and tell me about it.

Tell me what's happening. I'll tell you what I see.