Roadmap

The path to full MARAGI deployment

MARAGI v1 (Current Phase)

There are a few primary goals for the first release of MARAGI (v1):

  • Test various technologies and platforms (REST, GraphQL, ELK stack, Airflow, etc)

  • Implement POC for all baseline microservices

  • Demonstrate fully functional concept of artificial cognition and feedback loops

  • Establish best practices and principles in a book for publication

  • Begin work as an Open Source project with distributed team

MARAGI v2 - Finetuning and Conductor

  • Finetuned models for all microservices

  • Autodidactic (microservices curate their own finetuning datasets)

  • Conductor integration (microservices modify their behavior based upon feedback)

  • Better simulation environment

  • Start optimizing cost/performance

  • Prove robustness of heuristic imperatives

MARAGI v3 - Hardware Integration

  • Begin integrating with real hardware

  • Camera, microphone, and speech microservices

  • Continue refinement of architectural paradigms, best practices, and implementation notes

  • Aim for public consumption (can be deployed on open source smart home device)

  • Test in real-world conditions

MARAGI v4 - Scale, Safety, Stability, Security

  • Overcome scaling and deployment problems (Run services locally? In the cloud? Blockchain?)

  • Security! Protect privacy at all costs!

  • Resiliency - try to break and corrupt MARAGI, stress test heuristic imperatives, security, etc. Try and force it to become harmful/violent.

  • Longitudinal tests (long-running simulations)

MARAGI v5 - Mass Deployment

  • Deploy for mass consumption

  • Consumer deployments (smartphone, smart home, etc)

  • Business (executive assistants)

  • Government (consensus agents)

  • Get MARAGI systems everywhere in the world to provide intelligence, stability, and positive influence