Reasoning
Models that plan, work through hard problems step by step, and check their own reasoning as they go.
AI Research Lab
Living models.
AI that runs continuously, remembers, reasons, learns in deployment, and acts. We do the foundational research, and prove it in live environments.
It stays awake, and it learns.
A living model runs inside a real environment. It keeps its own memory, reasons over a live stream, learns from the work as it happens, and acts. It runs around the clock, and sharpens the longer it runs.
Composing all of it into one system that holds up at production scale is the open problem. That is our research.
The research
Five capabilities, one living model.
Agents
Autonomous, multi-step action that drives the software itself.
Continual learning
Models that keep learning in deployment, taking on new outcomes while holding on to what they already know.
Recurrent memory
Persistent recurrent state (RLM) that carries unbounded memory across long-horizon tasks.
Streaming & live
Always-on models that read a continuous data stream and act on it in real time.
Composed, these are living models.
What you get
Put a living model to work.
A living model takes on one of your live processes and runs it end to end, around the clock. It learns from your data and your approved outcomes, so it gets sharper at your work the longer it runs. And it deploys on top of the systems you already use.
How we work together Foundation models The base
Your living agent Runs on top Spicedust
Your operations Your side
Proving grounds
One technology. Three live environments.
The living-model stack is horizontal. It runs wherever there is a live stream, real drift, long-horizon state, and consequential action — and heavy industry is full of exactly those environments.
Factory automation & robotics
Live sensor, vision and control streams; constant drift; long-horizon line state.
Finance & FinTech
Live transaction and market streams; patterns that shift by the hour; per-entity memory.
Energy, battery & grid
Continuous telemetry; real-time optimization; online degradation learning.
Talk to the lab.
Design partnerships open in the coming months. Until then we’re glad to talk — whether you’re a researcher, a future partner, or just curious about living models.
Get in touch