Blog
Evidence-based notes on AI memory, coherence, and compliance.
- 24.04.2026·Dev Productivity·4 min read
Five Claude Code hooks that make coding agents actually reliable
Hooks are the layer where Claude Code goes from helpful suggestion engine to reliable coding partner. Here are five we run, what they catch, and why skipping any of them hurts.
- 23.04.2026·AI Memory·4 min read
Bi-temporal memory: why every fact needs a valid_to column
Most AI memory systems store facts as if they were true forever. Real facts have a validity window. Here is the pattern that fixes this — with a tiny schema change that unlocks a lot.
- 22.04.2026·Dev Productivity·5 min read
Giving Claude Code a memory that persists between sessions
Claude Code forgets everything when the session ends. Here is a concrete, working pattern for persistent memory - what to store, where to put it, and how to load it back in seconds.
- 21.04.2026·AI Memory·3 min read
Beyond RAG: why your AI needs episodic memory, not just retrieval
Retrieval-augmented generation is stateless by design. For agents that live longer than a single prompt, you need episodic memory - here is why, and how the shapes differ.