A cron job that nagged becomes a VM that finishes the job

Zeyliger built a small check to catch when his team's LLM gateway model list had drifted out of sync with a provider. The first version ran as a CI cron job that posted a Slack message telling someone to go fix it. He rebuilt it as a persistent virtual machine instead: the same check now proposes a diff, with a button to queue the change into the team's CI and CD pipeline, and the bot is checked into the same monorepo as the rest of the team's automated bots. "My old mindset was wanting a bot to nag me into doing the thing. The new mindset is to have the bot take the task as close to the finish line as possible," he writes.

Why it matters: The pattern moves the remaining human work from triage to review, a narrower and more specific ask than "someone should look at this." exe.dev sells the disposable-VM infrastructure this pattern runs on, and the post is a short personal account rather than a benchmarked comparison, so treat the underlying cost and reliability claims as unverified.