Driftless 2.0 Vision

In 2024, the AI industry convinced itself that the model was the product. Whoever had the best model won. Companies raced to put GPT-4, Claude, Gemini in front of users as fast as possible, slapping thin wrappers around frontier models and calling it a product. The pitch was simple: our model is smarter than their model.

That race is largely over — not because one model won, but because the models converged. The best models are, within narrow margins, interchangeable for most tasks. You can swap one for another and the average user won't notice. The moat everyone was building around model quality turned out to be a puddle.

The real race was never about the model. It was about continuity.

The thesis

Driftless was built on a single architectural bet: the model is a commodity, and continuity is the product. What users actually want — what they're willing to pay for, what they'll stay for — isn't a slightly smarter answer generator. It's an AI that knows them. One that remembers what happened last Tuesday, and last March, and three years ago when they made a decision that's still shaping how they operate today.

Driftless 2.0 is the build phase of that thesis. It's the platform we're constructing on top of the architectural decisions named in the first two papers — the Cooling Layer, the Sphere, the Gravity. Those aren't features. They're the foundation. Driftless 2.0 is what you build on that foundation when you decide to take the thesis seriously.

Bring your own model

One of the clearest expressions of the thesis is how Driftless handles AI models. You bring your own API key. You pick your model — Claude, GPT-4, Gemini, whatever you prefer or trust or can afford. Driftless doesn't mark up model costs or lock you into a single provider. The model is a commodity. We provide the continuity layer on top of it.

This is the move that most AI platforms get wrong. They bundle the model into their product because that's how they capture margin. But it creates lock-in, inflates costs, and bets the whole product on one provider's uptime and pricing decisions. Driftless bets on continuity instead — which is something no model provider can replicate or undercut, because it's built from your specific history with the system.

The platform shape

Driftless 2.0 is a platform, not a product. The distinction matters. A product is something you use. A platform is something you inhabit. When you join Driftless, you're not signing up for a subscription to a chat interface — you're setting up your Sphere, naming your Dottie, and beginning to build the Gravity that will compound in value for as long as you're here.

The crew — Vespera, Tuffy, Toodles, and others — comes with the platform. These aren't chatbots. They're functional agents with specific jobs: watching for freshness decay, tracking streaks, keeping the library current, flagging drift. They run automatically. They don't need to be configured. They exist because the platform is designed to run like an operation, not a chat window.

The surfaces are multiple: email, voice, web chat, mobile PWA. Each surface is appropriate for different moments — the email inbox for async and clients, voice for field work and quick capture, web chat for deep thinking sessions, mobile for everything in between. Your Gravity accumulates across all of them. Nothing lives in a silo.

Why now

The window for building this is open right now and won't stay open. The large platforms — OpenAI, Google, Anthropic — will eventually build continuity layers of their own. When they do, they'll have distribution advantages that are very hard to overcome. The question isn't whether they'll get there. It's whether Driftless can build deep enough relationships with enough users that switching becomes unthinkable.

A user whose Gravity contains three years of operational history doesn't switch to a new AI platform lightly. That's the moat. Not the model. Not the interface. The accumulated, irreplaceable, deeply personal context that makes their Dottie theirs — and makes starting over somewhere else feel like a real loss.

That's what we're building. That's the vision.

Filed by Rich Ligotino & Dottie — Driftless AI, 2026.