<?xml version="1.0" encoding="UTF-8"?><rss version="2.0"><channel><title>Learning out loud, Cherie Quek</title><description>Writing about how software gets built now that agents do more of the building.</description><link>https://cherie-q.github.io/</link><item><title>The next level needed software engineering</title><link>https://cherie-q.github.io/engineering-the-next-level/</link><guid isPermaLink="true">https://cherie-q.github.io/engineering-the-next-level/</guid><description>A desert-to-dessert typo helped my son imagine his own game worlds. Software engineering helped me keep building them: Everything vs. Zombies grew from 8 worlds to 22.</description><pubDate>Thu, 05 Nov 2026 00:00:00 GMT</pubDate></item><item><title>Building the process, not the feature</title><link>https://cherie-q.github.io/copilot-pr-squad/01-build-the-process/</link><guid isPermaLink="true">https://cherie-q.github.io/copilot-pr-squad/01-build-the-process/</guid><description>I stopped driving each Copilot PR by hand and built a team of agents that drives them for me. It got good the way a product does: watch it run, fix one thing, repeat. Building the process that builds the feature is now the job.</description><pubDate>Mon, 12 Oct 2026 00:00:00 GMT</pubDate><category>Building the process, not the feature</category></item><item><title>Watching it before trusting it</title><link>https://cherie-q.github.io/copilot-pr-squad/02-watching-before-trusting/</link><guid isPermaLink="true">https://cherie-q.github.io/copilot-pr-squad/02-watching-before-trusting/</guid><description>The squad wasn’t designed and then switched on. It got good through 96 small releases in eleven days, and it earned write access one step at a time.</description><pubDate>Mon, 12 Oct 2026 00:00:00 GMT</pubDate><category>Building the process, not the feature</category></item><item><title>Telling the loop what isn’t its problem</title><link>https://cherie-q.github.io/copilot-pr-squad/03-not-copilots-problem/</link><guid isPermaLink="true">https://cherie-q.github.io/copilot-pr-squad/03-not-copilots-problem/</guid><description>The most common way the squad wasted a round was asking Copilot to fix something that wasn’t Copilot’s to fix. Each case became a rule for recognizing it.</description><pubDate>Mon, 12 Oct 2026 00:00:00 GMT</pubDate><category>Building the process, not the feature</category></item><item><title>Catching scope creep early</title><link>https://cherie-q.github.io/copilot-pr-squad/04-catching-scope-creep/</link><guid isPermaLink="true">https://cherie-q.github.io/copilot-pr-squad/04-catching-scope-creep/</guid><description>An agent loop repeats a bad early choice every round. The fix was to catch scope as early as possible, and to make steering it a writing task.</description><pubDate>Mon, 12 Oct 2026 00:00:00 GMT</pubDate><category>Building the process, not the feature</category></item><item><title>Making it cheaper</title><link>https://cherie-q.github.io/copilot-pr-squad/05-making-it-cheaper/</link><guid isPermaLink="true">https://cherie-q.github.io/copilot-pr-squad/05-making-it-cheaper/</guid><description>Cost is easy to ignore until a busy day. Three small changes on Oct 7 cut the cost of a run by about a third, and the cost per PR on busy runs roughly in half. A cheaper model didn’t make the cut.</description><pubDate>Mon, 12 Oct 2026 00:00:00 GMT</pubDate><category>Building the process, not the feature</category></item><item><title>Lessons for your own loop</title><link>https://cherie-q.github.io/copilot-pr-squad/06-lessons/</link><guid isPermaLink="true">https://cherie-q.github.io/copilot-pr-squad/06-lessons/</guid><description>Most of what I learnt isn’t about pull requests. This is what I’d carry into building another squad, or any agent process that runs without someone watching.</description><pubDate>Mon, 12 Oct 2026 00:00:00 GMT</pubDate><category>Building the process, not the feature</category></item></channel></rss>