[crazylogger]

In a typical agent loop your N-th LLM request naturally becomes prefix for the (N+1)-th request. As the thread grows longer, cache hit rate converges to 100% and unit pricing for cached tokens is 10-100x cheaper.