MEMA vs. continual learning: three kinds of AI memory
Context windows, explicit MEMA records, and model-internal continual learning solve different problems. Here is the evidence boundary for MEMA, Nested Learning, and a planned HOPE integration.
Context Engineering
Don't build a company brain. Scope memory small.
Errors in memory don't repeat, they breed. So the size of a memory is the size of its risk. The fix isn't one brain that knows everything; it's many small, isolated memories, each scoped to one cohesive use case.
What an AI agent actually is
Everyone sells agents. Few say what one is. An agent is a model plus four decisions around it: what it sees right now, which rules always apply, what it remembers long-term, and where it runs. The memory part is the hard one.