Maybe it’s the content, maybe it’s me, but I’m noticing more of these lists trending towards software engineering practices, and fewer “new tools!” posts. I don’t mind that at all.
[blog] AI agent collisions: Worktrees aren’t enough. As you use more agents, separate git worktrees aren’t enough to prevent collisions. You still need explicit API contracts, short leases, and isolated resources before merging.
[blog] Scaling software built with AI: where the real bottlenecks are. Good look at where human attention should live in software now. It’s most about what to build, and proving that the output works.
[article] How will we make senior developers? No answers, but many have the same question.
[article] What behaviors do developers want from AI coding agents? We treat the agents like juniors, but don’t seem to invest as much in their improvement as we should.
[blog] Safe Vibe Coding in the Enterprise: Architecting a Governed Context Layer on Google Cloud. Excellent post, building on one I shared earlier this week. Proactively build out that context layer so that your builders are accessing the right data, the right way.
[blog] 2026: The State of Enterprise AI. Lots of charts and data points to convey how enterprise AI deployments are trending. And which companies are winning.
[blog] Agent Distraction. All those leftover TODO bug comments in your code are misdirecting your coding agents and making you waste time and tokens.
[blog] Inside GitHub’s new Git storage architecture. These look like some promising changes. They have time before all the competitors get marketshare.
[article] “Most apps never make it off localhost”: Cursor takes on Lovable and Replit with fully managed app deployment. Yes, the last mile matters. If you’re only coding and testing locally, you aren’t done. All these agentic coding tools, especially those from AI Labs, are missing a runtime. Expect partnerships!
Want to get this update sent to you every day? Subscribe to my RSS feed or subscribe via email below:
Leave a Reply