https://mokh.tech/lessons/tokenisation-the-cost-surface-of-llms/2026-01-12T09:00:00+00:00https://mokh.tech/lessons/forward-pass-what-actually-happens-when-an-llm-emits-one-token/2026-01-19T09:00:00+00:00https://mokh.tech/lessons/sampling-temperature-top-k-top-p/2026-01-26T09:00:00+00:00https://mokh.tech/lessons/kv-cache-the-serving-bottleneck/2026-02-02T09:00:00+00:00https://mokh.tech/lessons/chunking-the-decision-a-retriever-cannot-undo/2026-02-09T09:00:00+00:00https://mokh.tech/lessons/retrieval-bm25-vs-dense-vs-hybrid/2026-02-16T09:00:00+00:00https://mokh.tech/lessons/reranking-precision-purchased-with-latency/2026-02-23T09:00:00+00:00https://mokh.tech/lessons/parent-child-retrieve-small-serve-big/2026-03-02T09:00:00+00:00https://mokh.tech/lessons/diagnosis-retrieval-chunking-generation-failure/2026-03-09T09:00:00+00:00https://mokh.tech/lessons/the-twelve-components-an-llm-system-is-mostly-not-the-llm/2026-04-13T09:00:00+00:00https://mokh.tech/lessons/observability-the-trace-that-explains-a-decision/2026-04-20T09:00:00+00:00https://mokh.tech/lessons/cost-per-task-not-per-token/2026-04-27T09:00:00+00:00https://mokh.tech/lessons/indirect-prompt-injection-the-under-defended-attack/2026-05-04T09:00:00+00:00https://mokh.tech/lessons/workflow-vs-agent-decision-matrix/2025-12-01T09:00:00+00:00https://mokh.tech/lessons/bounded-agent-trace-playground/2025-12-08T09:00:00+00:00https://mokh.tech/lessons/agent-failure-mode-heatmap/2025-12-15T09:00:00+00:00https://mokh.tech/lessons/agent-safety-audit-the-layer-10-eval-gate/2025-12-22T09:00:00+00:00