A Paper Is Not a Workflow: Building an Auditable AI Review Skill
An auditable AI review workflow turns a paper into an evidence trail, staged critique, and a human-owned final judgment.
Read more →Hey, I'm
I write about AI, life, and the messy future of humans working with machines.
By day I'm a computer scientist at Lawrence Livermore National Lab building tools that help developers understand and optimize code. By night (and weekends) I write about AI agents, governance, intelligence, and anything else I can't stop thinking about.
An auditable AI review workflow turns a paper into an evidence trail, staged critique, and a human-owned final judgment.
Read more →
The problem was never knowing good principles. It was applying them consistently. AI agents change that equation β not just for individuals, but for nonprofi...
Read more →
What decades of software engineering taught me about running reliable AI agent infrastructure. Lessons from managing multiple AI gateways and why they need t...
Read more →
The smartest personal agent is often the most dangerous one, unless I split memory, execution, and approval into separate roles. I used to think the path to ...
Read more →
The Right Way to Use OpenClaw: Raise It Like a Partner, Not a Tool Most people underuse OpenClaw because they treat it like a smart utility; the real upside ...
Read more →
Sometimes I want to say βCodex, do the thing on my machineβ while Iβm away from my desk, on my phone, in a taxi, or between meetings. That was the whole reas...
Read more →29 articles across 6 themes spanning 6+ years of thinking.
Want to chat about AI, compilers, or something I wrote? Reach out.