Likely Flaws

We're pretty sure that several things will not be quite right.
Rules in the Soar and ACT are supposed to count for 50 ms of simulated time.
Our time fits right in with neurons, but rules have variable times for firing.
The binary CA is really an oversimplification of the way CAs work. They don't fire indefinitely. They can have variable activity.
Nothing is learned. ACT-R and Soar learn by chunking, that is making new rules.