Here's the handout for a three hour workshop I presented at the NICAR data journalism conference on using coding agents (Codex CLI, Claude Code etc) for data exploration, visualization and analysis
@simonw
-
Verify AI Answers with Working Code Evidence
By
–
The trick here is that you don't believe ChatGPT's answers to this kind of question until you've seen the code running yourself – the thing that's worth collecting is evidence that stuff works, not unproven answers from the bots
-
Agentic Engineering Patterns: Understanding Coding Agents Mechanics
By
–
New chapter for Agentic Engineering Patterns: I tried to distill key details of how coding agents work under the hood that are most useful to understand in order to use them effectively
-
Defining Agentic Engineering: Chapter 12 Guide
By
–
Just added the 12th chapter to my Agentic Engineering Patterns guide, but it's the first one in the sequence: I figured it was time to try and answer the obvious question, "What is agentic engineering?"
-
GPT-4 Vision: Image Processing in Large Language Models
By
–
They've been using the image as raw input to the model since GPT-4 vision back in late 2023 – here's a decent explanation of how that kind of model works
-
Dead Internet: AI-Generated Content Overtakes Authentic Discussion
By
–
Same here – it's gross, really dispiriting The time vampire slop replies outnumber the real ones here now I'm suspicious the same thing is starting to show up on Bluesky as well now Dead Internet.
-
Agentic Engineering: Fireside Chat at Pragmatic Summit
By
–
I spoke about agentic engineering at the Pragmatic Summit last month, in a fireside chat hosted by Eric Lui – here's the half hour video plus highlight quotes and extra notes from our conversation
-
Orchestrator Term Gaining Traction in AI Community
By
–
"Orchestrator" seems to be catching on for that from what I've seen
-
Liquid Template Language Performance Improved by 53%
By
–
Published some notes on @tobi
's autoresearch PR that improved the performance benchmark scores of the Liquid template language (which Tobi created for Shopify 20 years ago) by a hefty 53% https://
simonwillison.net/2026/Mar/13/li
quid/
… -
Small AI models pose greater rogue risk than larger ones
By
–
My intuition is that small models are far more likely to go rogue than big ones