Introducing Ruliad
A frontier AI research lab pursuing superintelligence beyond the transformer paradigm.
Nov 27, 2024
We are Ruliad, a frontier AI research lab pursuing superintelligence.
We believe the next breakthrough will not come from scaling LLMs further, but from discovering new architectures, new training methods, and continually learning systems beyond the transformer paradigm.
What we believe
The current paradigm has reached its limits. Superintelligence will not come from squeezing more out of static LLMs. It will come from new architectures, continual learning, fluid systems, and novel training methods. We are exploring the tech tree of possible paths beyond transformers, identifying and building the next paradigms as they emerge.
Insight
We believe convergent evolution of intelligent agents in nature is evidence that intelligence can be engineered, not just imitated. Nature shows that intelligence emerges through adaptive, convergent processes, and we believe digital systems can do the same. This is the moment for an AI Cambrian explosion. Not one path, but many. We are here to find the ones that matter.
Who we are
We use AI to accelerate our own research and build advanced systems for enterprises and governments for use in real world applications.
We are a small team. Our infrastructure is largely automated by the systems we build. Self-improving pipelines that serve as a live proof-of-concept of our own thesis. Smaller labs can matter precisely because constraints force invention.
What to expect
We will share work we find interesting: papers, models, technical essays, and results, whether or not they fit the current narrative.