Defining The Future Of AI
- Eros
- Sagas
- Orion
- Self Driven Systematic Growth
- Ouah
- Athena
- Genesis
- Odysseus
- Icarus
- Argus
- CNA
- Eos
- Dynamic learning systems
- Ariadne
- Dynamic Memory
We can only see a short distance ahead, but we can see plenty there that needs to be done.
Curious about our research? Just ask.
From our discoveries to the work behind them, if it’s about Forlais, you’ll find it here.
Forlais is at the frontier of original AI research spanning intelligence, learning, memory, autonomy and computing.
Each programme represents a unique intelligence layer designed with distinct purpose. Each layer is the foundation upon which the next one sits, creating a tailored intelligence system that fits perfectly together.
Latest Research
What our papers and experiments have found about trained models in wide use today.
Finding · 27 June 2026
The Forlais Theorem
For trained models as they are commonly stored and shared, there is a firm limit to how far conventional exact compression can go, and it can be stated before any compression is attempted.
Finding · 27 June 2026
The Entropy Wall
Conventional exact compression of trained models stops at a limit, and the limit is shared across the model families we measured.
Finding · 27 June 2026
The Mi Paradox
A signal in trained models that appeared to offer a route to better compression turned out to be largely an artefact of measurement.
Finding · 27 June 2026
The r=2 Neural Law
Trained models from several families and developers share the same regularity.
Research areas
Each area has its own account: the problem, what we set out to do, what we found and where the work stands.
- CNA Architecture A floatless model architecture with compression in its foundation.
- Memory Memory that keeps what a model knows, recovers it exactly and can be relied on.
- Dynamic Memory Research into memory that learns from the work as it happens.
- Compression Trained models in less storage, with every part restored exactly.
Programmes
Each programme investigates a different part of the wider problem. Together, they create the technical layers from which Forlais can build complete AI systems.
The Research Index- Orion Exploring new foundations for computation.
- Genesis A different foundation for intelligence.
- Self Driven Systematic Growth Discovering patterns. Developing understanding.
- Athena Memory that can remain, recover and develop.
- Ouah Preserving what a model carries forward.
- Ariadne Bringing memory back into a model's work.
- Odysseus Learning relationships. Applying them to something new.
- Eros Learning, retaining and developing through experience.
- Sagas Bringing the abilities of a model together.
- Argus Sustaining observation. Continuing investigation.
- Icarus Developing new abilities. Preserving what has been learnt.
- Eos Discovering relationships. Carrying understanding forward.
- Dynamic learning systems Learning that changes what becomes possible next.