Glenn Thimmes

Async collaboration isn't more efficient collaboration. It's less collaboration. Which is fine, but stop calling it teamwork and call it parallel work with occasional check-ins. Collaboration takes time or it's a Slack thread people thumbs up without reading.

Retrospectives without follow-through are just therapy sessions where nothing changes. Cathartic? Maybe. Valuable? No.

Your test coverage percentage is a vanity metric. How fast do you know if you release a bug into the wild?

On the Noble Uses of AI When Cognitive Offloading Elevates Us

AI should free us from the mechanical so we can focus on the meaningful — not replace the hard work of thinking itself. A must-read for anyone thinking about why and how we integrate AI into our lives.
blog.cosmos-institute.org/p/on-the-nob...

Unpopular opinion: The standup isn't broken. Your team's psychological safety is. When people feel safe, updates flow naturally.

Startup velocity isn't about working faster. It's about making fewer wrong decisions before running out of runway.

Unpopular opinion: "move fast and break things" is fine for features.

It's terrible advice for team trust.

Your architecture diagrams are lies. Useful lies. But lies. The real system is in the incident reports.

A lesson that some leaders get right away and others never learn: your job isn't to have the best ideas. It's to create an environment where the best ideas win — even when they're not yours.
#Leadership #SoftwareEngineering

YouTube video by 3Blue1Brown Transformers, the tech behind LLMs | Deep Learning Chapter 5

This is one of the best explanations I have seen of the tech behind LLMs. If you have 27 minutes to spare, you'll find it surprisingly easy to follow.
www.youtube.com/watch?v=wjZo...

Aptitude assessments up 54x since 2024. When AI generates code, the bottleneck moves upstream. Syntax depreciates. Judgment doesn't.
buff.ly/Tk6ZlCN #SoftwareEngineering #TechHiring #EngineeringLeadership

Generative coding: 10 Breakthrough Technologies 2026 AI coding tools are rapidly changing how we produce software, and the industry is embracing it—perhaps at the expense of entry-level coding jobs.

AI is writing 30% of Microsoft's code and 25% of Google's. How we measure productivity is more important than ever. The teams that win will measure outcomes: reliability, time-to-ship, customer impact. AI is forcing us to become better leaders. #ai
www.technologyreview.com/2026/01/12/1...