https://blog.oklensuangoo.uk/posts/Recent content in Posts on Architecting IntelligenceHugoen-usSun, 14 Jun 2026 01:00:00 +0800https://blog.oklensuangoo.uk/posts/kv-cache-remembers-deleted-context/Sun, 14 Jun 2026 01:00:00 +0800https://blog.oklensuangoo.uk/posts/kv-cache-remembers-deleted-context/Reusing a subset of a KV cache — training-free — is often more accurate than recomputing from the kept text. This post gives the variable that governs it (coverage), the causal mechanism behind it (a downstream-attention trace), and what it means for agent context compression.https://blog.oklensuangoo.uk/posts/agent-memory-evolution/Sun, 15 Mar 2026 17:47:02 +0800https://blog.oklensuangoo.uk/posts/agent-memory-evolution/As LLMs turn from language models into the brains of agents, persistent context management becomes the new challenge for RAG and memory. A survey of token-level / parametric / latent memory paradigms and the emerging benchmarks (Memory Arena, Evo-Memory / ReMem).