Memory Agent Review We created a chatbot that saves semantic memories to a single user profile or collection. We introduced Trustcall as a way to update either schema. Goals Now, we're going to pull together the pieces we've learned to build an agent with long-term memory. Our agent, task_mAIstro, will help us manage a ToDo list! The chatbots we built previously *always* reflected on the conversation and saved memories. task_mAIstro will decide *when* to save memories (items to our ToDo list). The chatbots we built previously always saved one type of memory, a profile or a collection. task_mAIstro can decide to save to either a user profile or a collection of ToDo items. In addition semantic memory, task_mAIstro also will manage procedural memory. This allows the user to update their preferences for creating ToDo items.
Autonomy
Semi-autonomous
Sandbox-aware
No declared sandbox guidance
Network access
Unspecified
Filesystem access
Unspecified
Permissions declared
Not declared
Pattern
Single agent
Models
gpt-4oclaude-3-5-sonnetgpt-3.5-turbo