Category: Note

When this system is up and running, because it can learn continuously (unlike current AI), it will be just like Johnny 5 from Short Circuit movie. More input!

I see the current AI as the next evolution of search.. It brings search into the context where i’m working.. therefore it can build more context around what i’m searching. It uses a different kind of index than google, far more semantic.. It isn’t intelligent in many ways but certainly much smarter than “google search”.…

Watching “2025/09 – How Does the Neocortex Split and Combine Models Without Relearning”.. I feel like all the confusion is arising from the fact that Monty’s implementation is too software-like.. Poses, for example, have to use grid-cells, or grid-cell-like implementation, not 3D vectors or rotation matrices.. Raw sensor data has to be converted to proper…

Based on the team discussions I’m watching in the more recent videos, there is a disconnect between how @JeffCHawkins is thinking and how the Monty folks are thinking about the problem. Jeff is approaching it from the neuroscience point of view, where the brain is an analog computer with zero domain knowledge and has no…

Just watching “2025/07 – Hierarchy or Heterarchy”.. It’s partially talking about the parent/child object issue. ChatGPT and I thought those designations aren’t a good idea as it would break the “all the columns work the same way” principle. Columns shouldn’t care about objects, or their composition.. we talk about them from our POV, but that’s…

“2025/07 – Proposal for Separating Behavior Models” proposes breaking up morphology modeling and behavior modeling from being inside a single LM to separate LMs. Two ways to look at it: In terms of software/Monty, these are minor changes. In terms of the neuroscience, Jeff says Monty doesn’t have to be a neurosimulator. His red line…

One thing that has been a continuous source of confusion in understanding how composite objects are recognized and how behavioral models are built (which necessitate multiple objects) is: I haven’t caught up with the latest videos yet, so perhaps these will be resolved when I catch up.. or maybe not..😬 Given: Trying to explain and…

This video is titled “A Solution for Behavior in the Cortex” but it is a great overview of TBT and how everything fits together There were several confusing bits about TBT that I think are finally sinking in after re-watching the videos (in chronological order) and also conversing with chatgpt. It’s amazing the low level…

One thing I initially missed about TBT is, when I heard it’s a “sensorimotor” framework, I started thinking about robots, but it really is talking more generally about how “locomotion”/movement is crucial to learning and intelligence. Current AI agrees with TBT on:1) Features: stable patterns that are shared between many inputs Then there is:2) Features…

DNNs create features and combines them to recreate the input data during training. Both the features and the percentage of which feature adds up to which input is stored in the same parameters. Therefore every new training data has the potential to mess up previous representations of data. Monty has a more sophisticated learning algorithm…