If your AI coding agent ships a full feature in 1–2 giant commits, debugging production issues becomes a scavenger hunt. Require atomic commits: one migration, one handler, one test update. Small, self-contained changes create a usable audit trail. Save this rule.
Tips to be more productive with AI:
1. Prioritize concurrency.
2. Break work into clear lanes for AI agents.
3. Launch those lanes early while you tackle the human work.
4. Find the manual bottlenecks—they’re where your attention creates the most leverage.
#AI#Productivity
Github seems to have an issue affecting the commit graph for their top contributors.
Including @steipete and @yegor256
Perhaps they need to be re-indexed?
#github#commitgraph
AI news is not your to-do list. New releases are worth tracking. They are not a reason to rebuild your workflow every week. Pick a stack. Learn it deeply. Ship something real. Save this for when the AI feed makes you feel behind.
@robj3d3@AnthropicAI I get it. Riding the wave of releases is like chasing the best wave. Each time they market it as ground breaking, each time I spend hours testing late nights and early mornings. Let’s all stop and take a breath.
One thing I think developers and vibe coders underestimate is the agent's.md. Everyone is obsessing over prompts. Meanwhile, I'm seeing more leverage from spending 30 minutes improving... #rogerchappel#roger
One thing I'm experimenting with is giving coding agents work while I sleep. Not generating code always, but reviewing repositories, finding dead code, updating dependencies, finding... #rogerchappel#roger