reading the EP2 book I came across the Memory Lock on page 320.
Memory Lock: When activated, this implant prevents your sensory input from being stored in your long-term memory, tagged by mnemonics, or recorded/transmitted by your mesh inserts or other means. It also temporarily blocks cortical stack backups. You retain short-term memories, but for no more than a few minutes. This implant is often a requirement of personal aides, consultants, couriers, and underlings of powerful people who require confidentiality or deniability.
this means the process of storing long term memories is understood enough to interrupt without affecting short term memory, personality or reasoning. in effect we have the address for the long term memory system. so what is to stop us from calling it from multiple processes, like a central database.
my idea was to have multiple forks of a single ego all reading and writing long term memories to the same ego.
some code to explain
this would appear as multiple forks of the same ego behaving independently with the exception that they become aware of what the others are doing as long term memories are written, generally a couple of minutes after the fact. in effect, you can be in multiple places doing multiple things without a significant problem of cross fork awareness.
from a game point of view, this can not be allowed to work. it breaks the action economy and removes most of the downsides of forking. I was, however, thinking of it as having potential as an ex-human strategy if I can come up with suitable downsides to explain why only exhumans would use it.