Yoonbo Cho — writing https://yoonbocho.com Native mobile apps and applied LLM systems. en-us HIPAA and LLMs in late 2026: what a BAA doesn't cover https://yoonbocho.com/writing/hipaa-and-llms-what-a-baa-does-not-cover/ https://yoonbocho.com/writing/hipaa-and-llms-what-a-baa-does-not-cover/ Mon, 02 Nov 2026 12:00:00 GMT Notes for engineers putting health data near a model: BAA scope, zero data retention, ePHI you forget you have, and the state laws that apply anyway. An eval harness for a one-person AI product https://yoonbocho.com/writing/an-eval-harness-for-a-one-person-ai-product/ https://yoonbocho.com/writing/an-eval-harness-for-a-one-person-ai-product/ Sun, 01 Nov 2026 12:00:00 GMT Release bars for a nondeterministic product when you are the whole team. A 14B local model beat a 26B one on my real task https://yoonbocho.com/writing/a-14b-local-model-beat-a-26b-one/ https://yoonbocho.com/writing/a-14b-local-model-beat-a-26b-one/ Fri, 30 Oct 2026 12:00:00 GMT Choosing a local extraction model by measuring two candidates on the same real pages. Don't let your app promise more privacy than your config delivers https://yoonbocho.com/writing/privacy-copy-should-match-your-config/ https://yoonbocho.com/writing/privacy-copy-should-match-your-config/ Wed, 28 Oct 2026 12:00:00 GMT How Everblue decides what it may say about AI and your data: on-device first, zero data retention as a flag, and a never-claim list. What I learned de-identifying clinical documents for an LLM https://yoonbocho.com/writing/deidentifying-clinical-documents-for-an-llm/ https://yoonbocho.com/writing/deidentifying-clinical-documents-for-an-llm/ Mon, 26 Oct 2026 12:00:00 GMT Notes from Arbiter, a clinical-document RAG prototype on synthetic records: Safe Harbor, flattering recall metrics, and where Presidio's defaults miss.