Notes on building useful AI.
Practical notes on models, tools, experiments, and engineering. We share project lessons and technical reading for people who build and buy AI systems.
Turning architectural plans into a browser walkthrough
Lessons from a browser 3D walkthrough: a rebuildable Blender model, baked lighting, streamed assets, performance checks, and a guide using plan sources.
All notes
A voice agent’s rules can’t live in its prompt
Use source tools, server-side action checks, and session limits to control voice agents. Prompts alone cannot enforce permissions or spending.
Context engineering for coding agents
Give coding agents clear repository instructions, working checks, and useful examples. Make each task easier to implement and review.
Choosing a model is an eval problem
Compare AI models on your own tasks. Measure answer quality, cost per task, and response time, then test routing and fallback options.
Evals start with real tasks
Build AI evaluations from real tasks, define expected outcomes, and test critical failures. Use the results to guide release and ongoing review.
A grounded answer shows its source
Build knowledge agents with traceable sources, version details, access controls, and clear responses when information is missing or conflicts.
Self-improving agents need evidence loops
Turn expert corrections into reviewed evidence, targeted evaluations, and software changes. A practical approach to improving AI agents after launch.
Sandboxing agents requires real enforcement
How to assess a coding-agent sandbox: file access, network rules, child processes, and the actions that still need approval.
Voice AI latency depends on more than the model
How network routing, session setup, and interruption handling affect voice AI. Engineering lessons from OpenAI’s real-time media infrastructure.
Get new notes
Tell us what needs to work.
Describe the work you want to improve.