Taking seriously 'what if the AI gets it wrong'
With reference to SNAP as an illustration
Plausible Legibility
Systems (technical and institutional), practical knowledge, and a bit of free play.
Subscribe:.rss.atom.json.md.m3u.pls
Live Last read · last published · next check
Latest episodes
Working in the open is much more valuable than I once thought
It’s time to mark to market a view of mine that has evolved over the past decade.
Diagnosing government incapacity: alignment, dysregulation, capability
Now seems like a moment of reflection who work on some form of “improving government.” I see a lot of people wrestling with ideas and theories of change and where we are now.
Building a real-time state update pipeline for 370,000 SNAP recipients with AI
This is a crosspost of a write-up I did on the Propel blog about how we implemented a pipeline of monitoring the 50 state SNAP agencies’ websites for updates on the government shutdown, and piped it out to ~370,000 SNAP participants via the Propel app.
We need to start talking about an AI benefit design
And more specifically: how we design it to actually work
AI models are getting dramatically better at complex policy questions
Evidence from SNAP Asset Limits
Kernels 02: Claude Compadre (Claude Code for non-coding things)
Also: is curiosity all you need?
Kernels #01
Ideas I'm wrestling with (a first edition)
Edge cases as policy reality: using AI to navigate SNAP purchase restrictions
*and* to front load the intrinsic complexity of implementation
Little-t tools for thought
On folk software, plumber programming, and agency from making things that work for you