# git rebase main topic
$ git checkout topic
$ git rebase main

%%{init: { 'logLevel': 'debug',
    'theme': 'neutral',
    'gitGraph': {'rotateCommitLabel': false},
    'themeVariables': {
              'commitLabelFontSize': '14px'
       } } }%%
    gitGraph LR:
       commit id: "C1"
       commit id: "C2"
       branch "old topic" order: 3
       commit id: "C3"
       commit id: "C4" tag: "ORIG_HEAD"
       checkout main
       commit id: "C5"
       commit id: "C6"
       branch "topic"
       commit id: "C3′"
       commit id: "C4′" tag: "HEAD"

Links to articles I’ve found useful as I figure this LLM stuff out.

A Python program that uses the ChatGPT API to generate a complete web app by chaining multiple prompts and using intermediate results.

• Reshaping JSON with jq | Programming Historian
• Introducing zq | Brim Data

Some notes on a mechanism for posting from Drafts with as little friction as possible.

I put together a repo of some GPT toys for
Drafts that I’ve been playing with.

I want this repo to only ever have three commits.

Dot notation lets you specify sets of commits.

Do you find rebasing in Git confusing? I do.

Tables are hard to work with in Markdown. Here’s an easier way that uses lists.

When I was figuring out how collections work, I tried my hand at implementing a collection of collections. I called the things in the collection of collections categories.

Collections are a key concept of Eleventy, but they confused me at first. So I looked at the source to figure out how collections work.