Memory & recovery
Pull the plug.
Two systems. The same task. One keeps what it learns. Interrupt them and see what carries on.
Try the experimentWe pursue our own research into artificial life, persistent agents, memory and adaptation, alongside research commissioned for specific technical questions.
[ Try it ]
What survives a restart? What can continue without a connection? Explore decisions that shape the systems we build.
Memory & recovery
Two systems. The same task. One keeps what it learns. Interrupt them and see what carries on.
Try the experimentData & connection
A field report. Three ways to process it. Choose where the work happens, then see what changes without a connection.
Try the experimentHow can a system adapt to new conditions while preserving the behavior an organization relies on?
The investigation
Controlled variation and feedback offer ways to adapt software to changing workloads. We examine how to retain useful behavior and detect regressions during that change.
What must an agent retain to resume an operational task after an interruption or handover?
The investigation
We investigate the task state and tool coordination needed for work that outlasts a session, including interruption, handover, and resumption.
How should a system use operational history when conditions change?
The investigation
Simulated environments let us examine how sequences of observations inform decisions, and when a changing situation makes earlier information less useful.
How can an AI system carry useful context forward while keeping that context understandable and controllable?
The investigation
We explore memory and interaction designs that retain useful working context across repeated use, while letting people inspect, correct, and control what persists.
[ Practical Application ]
Our independent research explores artificial life, persistent agents, memory and adaptive systems. We investigate how these systems develop, retain useful experience and continue working over time.
Research gives us new approaches to test in practical work. Engineering and deployment bring their own constraints and unresolved questions back to the lab. You can also commission research around a particular problem, either on its own or as part of a larger assignment.
How a research engagement develops[ Research & Collaboration ]
Share the behavior you need to investigate and the decision the research should inform.