Context management for AI systems
noteCM-007
Either the agent writes during the turn and the user waits, or it writes afterwards and the memory is not there for the next question. Each choice hands you a different set of problems.
guideCM-006
Facts, events, and how-to. The split is borrowed from human-memory research, and two public sources already disagree about what belongs in the third category.
noteCM-005
The model remembers nothing between calls. The context window is the working area for one call, not a store — so anything that persists lives outside the model and has to be re-supplied.
guideCM-001
No. The study everyone cites says something much narrower, and four of the field's other most-repeated statistics do not survive contact with their sources either.
Want the agent reading your mail to know how to better manage its context? Subscribe.
About
Ori is a founder, keynote speaker, and AI consultant. With more than a decade of experience working with data — from data science to agentic systems and the infrastructure they run on — he deeply understands how data behaves in the wild, and how to harness it to build intelligent, scalable solutions.