Geometric AI Memory

Curved neural networks enable AI memory recall through geometric design

The critical take-out from this is they discovered geometry is an intrinsic aspect of memory, and how we can make memory work more efficiently. This is the same as my own work, and has implications for quantum computing. Already, the folds in our brains have been correlated to the individual's personality and, in a Singularity, everything inevitably transforms into geometry. Combining the right geometry with a collection of pattern matching neural networks, empowers them to sometimes be 125% efficient.

Efficiency has no demonstrable meaning, outside of a specific context, and every context must have its own geometry.

 

They should try using bubble memory.
Everything is in a bubble.