Hi everyone,
It’s hard to believe that it’s been two years since the Thousand Brains Project began building an open-source platform for a different approach to AI, using a sensorimotor framework inspired by the human neocortex.
Today, we’re sharing our Two-Year Report. It takes a detailed look at what has worked, what we’ve learned along the way, what Monty can do today, and the problems we’re excited to tackle next.
Read the report here.
In the past two years, we’ve shipped 54 public releases on GitHub, reaching v0.47.0 in August 2026. A lot of our work on the engineering side has been to transform Monty from a codebase used primarily by our researchers to a platform that people outside our team can use, extend, and build on.
And it hasn’t just been us alone. 16 outside contributors have authored 121 merged pull requests, helping to enhance everything from Monty’s architecture to implementing new features and improving our documentation. We are so grateful for all the help we’ve received from our incredible community.
On the research side, one of the biggest questions we’ve explored has been how Monty can understand a world made up of objects composed of other objects. Today, Monty can now combine representations of smaller objects into larger compositional models: for example, recognizing both a 3D object and a 2D logo on its surface and learning how they are arranged relative to each other. This is a crucial requirement for an intelligent system to be able to efficiently and robustly model the compositional world we live in. Introducing this new ability into Monty required several other advances, such as dynamically resampling hypotheses to deal with a more complex world, intelligent policies that quickly move to salient regions, lower-dimensional sensor modules to handle object distortions, and an attention system for Monty.
We’ve also been testing Monty outside the simulator, working with real motors and various sensors, including ultrasound.
There are still many questions left to solve. But after two years, the problems facing us are much clearer. And so is our path forward.
You, the members of our community, are a huge part of the Thousand Brains Project. If any part of this report catches your interest, let us know!
With excitement,
The Thousand Brains Project Team
