Knowledge Gaps are the concepts the AI stream keeps surfacing while they remain under-understood — the radar’s signal that a topic is worth deliberate learning time, not just passive reading. Each gap names why it is showing up now and links the concepts that make up a path to close it.
They are the raw input to what AI builders should learn next: the current gaps cluster around the retrieval stack (RAG, Embeddings, Reranking) and the agent stack (Agent Memory, Context Engineering, Codebase Indexing).
The way to use a gap is the way the radar frames it: pick one, build a short learning path from its linked concepts, and apply it to one real workflow of your own. Browse the open gaps below.