RGI Labs' growth trajectory would change if its research program delivered a demonstrated working system rather than stated ambitions; a published result or a product surface that turns its contact invitation into work would confirm that change. The program pursues recursive and differentiable self-improvement across architecture, parameters, and scaffolding, alongside continual learning from online experiences through parameter updates and long-term memory.
RGI Labs invites readers to contact the company to learn more about the company's work.
2Memo
The investment case for RGI Labs is a one-person lab pursuing general intelligence through online learning, world modeling, and recursive self-improvement; it holds only if that research program produces a working system and a route from its contact invitation into a real offering. Founded in 2021, part of Y Combinator's Summer 2021 batch, and based in New York City with a team of one, the company also holds that reaching AGI requires multi-step planning, long-term memory, few-shot skill acquisition, and latent space for reasoning.
RGI Labs invites readers to reach out by email to learn more about the company's work.
3Risk Factors
Concentration is the risk that could break RGI Labs' thesis: the company is a one-person team still listed as active in New York City whose outward surface is an invitation to make contact, so a stumble in the program would leave no second person, product, or buyer to absorb it.
RGI Labs pursues two research goals: using a world model to predict future states and plan through multi-step rollout, and continual learning from online experiences with parameter updates and long-term memory.