You might remember the pre-print we shared last July, describing Monty’s capabilities and comparing it to transformers on a variety of measures (data efficiency, compute efficiency, accuracy, …). We’re excited to announce that the paper is now published in the journal Neural Computation 
You can find the peer-reviewed version here:
For a condensed, plain language explanation and video of @nleadholm talking you through the many exciting results, check out our website: Thousand-Brain Systems: Sensorimotor Intelligence for Rapid, Robust Learning and Inference: A Plain-Language Explainer | Thousand Brains Project
