Frontiers
Three frontiers. One research thesis.
Model Architecture, Applied AI for Businesses, and Personalised AI for Everyday Life. Aurora research validates hypotheses in controlled environments. Frontiers 2 and 3 apply those results in products — using existing LLMs, layered with Aurora-derived memory and event systems.
Model Architecture
Intelligence should be organised around sparse events, local predictive computation, and multi-timescale memory — not around massive dense matrices trained once at enormous cost.
Read the frontier brief →What we ruled out
- Building or fine-tuning a frontier LLM
- Transformer variants or token-prediction optimisation
- Standard RAG systems or retrieval wrappers
- Scaling existing architectures as the primary research direction
Projects in this frontier
Applied AI for Businesses
AI becomes genuinely useful to a business when it is embedded in a real workflow, handles the structured majority of decisions autonomously, and routes the nuanced minority to humans with proper context.
Read the frontier brief →What we ruled out
- Hallucination-tolerant consumer apps
- Generative ad-tech
- RAG over toy corpora
- Operating infrastructure on the customer's behalf
- Consulting engagements that produce decks instead of shipped software
Projects in this frontier
Personalised AI for Everyday Life
Re-explaining your life to an AI every session is evidence that the system is not paying attention. Personalised AI should learn continuously, act proactively, and never leave your device.
Read the frontier brief →What we ruled out
- Cloud-default context storage
- Always-on listening packaged as "context"
- Engagement metrics in personal AI products
- Sending personal context to a hyperscaler without explicit opt-in
Projects in this frontier