Writing
Blogs on building, learning and product thinking
- Sep 30, 20266 min readNew
Your AI Didn't Lie to You. It Was Never Trying to Tell the Truth.
Every person who has used an LLM has been confidently handed something that does not exist. Here is why it happens, in plain language, and what actually fixes it.
Read article
- Jul 5, 20265 min read
The Most Dangerous Line of Code Might Be `retry()`
Retries look harmless from one client, but at scale they can synchronize traffic and make recovery harder. The missing ingredient is jitter.
Read article
- May 29, 20266 min read
Underneath the abstractions
The smoother software feels on the surface, the more invisible work is happening underneath.
Read article
- Apr 18, 20266 min read
The Building Blocks of an AI Agent
A clear mental model of AI agents: how models, tools, memory, and loops come together to solve complex tasks step by step.
Read article
4 articles