Tag
A detailed technical explanation of how CPU caches work, covering the principle of locality, cache organization, indexing, and handling writes.
This paper proposes a learnable Dirichlet-process cache that allocates memory slots only for novel inputs, enabling efficient associative recall with a cache size proportional to distinct items rather than token count. It combines DP-means clustering with recurrent backbones, demonstrating effectiveness on associative recall benchmarks and real-world streams.
Cloudflare launches Workers Cache, a tiered cache that sits in front of Workers, allowing cached responses to be served without invoking the Worker, reducing CPU time and improving performance.
Maka's Harness project improved the self-check mechanism, enabling DeepSeek Flash V4 to achieve evaluation results close to GLM-5.2 on the terminal-bench sample set, completing 10 programming agent tasks with only 4 RMB and a 97.5% cache hit rate.
The author argues for using memcached over Redis as a caching layer, highlighting its simplicity, ease of handling downtime, and straightforward clustering, contrasting with Redis's feature creep and tendency to be misused as a persistent database.
Anthropic AI engineer demonstrates a free 4-step method to give AI agents persistent memory across sessions, including memory stores and dreaming, achieving 95% cache hit rate.