About us
This group brings together researchers, engineers, founders, and technical leaders exploring recursively self-improving AI, metareasoning, and AI neofactories.
Hosted by Sanscritic, a neolab building metareasoning systems for expert work, the group focuses on AI systems that can evaluate their own reasoning, resolve conflicting perspectives, learn from a small number of expert questions, and improve over time.
We will host technical talks, research discussions, paper clubs, demos, and founder presentations covering topics such as:
- Recursive self-improvement
- Metareasoning and reasoning architectures
- Open-weight models
- Self-evaluating AI systems
- Simulation-driven AI
- AI neofactories
- Applications in enterprise, robotics, biotech, science, and deeptech
Our core belief: in expert work, reasoning is not just a step toward the answer. Reasoning is the product.
Join us to meet others building and researching the next generation of intelligent systems.
Past events
2
