cluster-cache-no-recompute
IN premise
`ClusterCache.embed()` does not recompute embeddings for previously cached belief texts; cache size stays constant on repeated calls with the same input and grows by exactly the count of new texts on superset calls.
Summary
When the embedding cache receives texts it has already processed, it skips the expensive embedding computation and returns the cached result. This means repeated or overlapping calls are cheap and predictable — the cache only does work for genuinely new inputs, so performance scales with novel content rather than total call volume.
Details
| Source | entries/2026/05/08/tests-test_cluster.md |