// writing
Posts tagged #ai
A mix of technical deep-dives, career reflections, and dispatches from the road.
tag: #ai
AINov 8, 2026· 4 min
LLM Rate Limits: Your Retry Loop Is the Easy Part
Not every 429 is worth retrying. How to handle LLM API rate limits without stacking retries, burning budget on spend-cap errors, or finding limits by failing.
read more →
AINov 6, 2026· 4 min
Writing an MCP Server: The Tool Description Is the Whole API
A walkthrough of building an MCP server with the v2 TypeScript SDK, and why your schemas, descriptions, and error messages matter more than protocol plumbing.
read more →
AIOct 31, 2026· 3 min
LLM Evals Start as Thirty Test Cases, Not a Platform
Most LLM features ship on vibes and break on the next prompt tweak. A small regression set, cheap deterministic checks, and a calibrated judge catch most of it.
read more →
AIOct 30, 2026· 4 min
Prompt Caching: Your Prompt's Prefix Is a Cache Key
Prompt caching can cut repeated input costs by around 80%, but only if the stable parts of your prompt come first. Here's how it works and what breaks it.
read more →
AIOct 24, 2026· 4 min
Prompt Injection Is an Architecture Problem, Not a Prompt Problem
Why system-prompt warnings don't stop prompt injection, and the lethal trifecta, taint tracking, and quarantined-LLM patterns that actually constrain an agent.
read more →
AIOct 23, 2026· 4 min
Structured Outputs: Stop Parsing JSON Out of LLM Prose
Constrained decoding means the model can't return malformed JSON. Here's how structured outputs work with Zod in TypeScript, and the edge cases they miss.
read more →
// the newsletter