Developers Don’t Need More Documentation
- Dennis Pilarinos tl;dr: Docs get written, but answers stay hard to find. The problem isn’t the docs themselves. It’s that the context developers need is scattered, outdated, or missing entirely. Why does this keep happening? And what’s the alternative?featured in #617
Why DIY Internal Tools Collapse — And What CIOs & CTOs Should Do Instead
- Gibbs Cullen tl;dr: Vibe coding tools like Lovable, Bolt, and Replit promise speed—but often lead to shadow IT, tech debt, and security risks. As AI accelerates business demands, brittle DIY foundations can’t keep up. The answer? A modular, centrally governed approach built for scale, extensibility, and enterprise-grade AI. Superblocks has collected insights from engineering leaders at top enterprises into this blog to help you understand the risks—and how companies are preparing for the future of internal tooling. Read more to find out how you can too.featured in #617
GitHub’s Engineering System Success Playbook
tl;dr: “Inspired by multiple frameworks, including SPACE and DevEx, DX Core 4, and DORA, our playbook offers a balanced and comprehensive approach, helping you assign metrics to each “zone” that you can track over time and iterate as needed.”featured in #617
How Do AI Code Reviews Impact Engineering Teams?
- Lizzie Matusov tl;dr: “Researchers conducted a study inside the software division of Beko, a multinational electronics company. Over 10 months, 238 developers used an AI-based code review tool powered by GPT-4. The study analyzed 4,335 pull requests across three projects, comparing automated vs. manual review patterns, survey feedback, and developer behavior.”featured in #616
Innovations In Evaluating AI Agent Performance
- Michael Lopp tl;dr: Just like athletes need more than one drill to win a competition, AI agents require consistent training based on real-world performance metrics to excel in their role. At QA Wolf, we’ve developed weighted “gym scenarios” to simulate real-world challenges and track their progress over time. How does our AI use these metrics to improve our accuracy continuously?featured in #616
Things I Currently Believe About AI And Tech Employment
- Camille Fournier tl;dr: “This is not based much on the current state of innovation, except to take as a given that AI is proving to be measurably supportive of building software. It's instead based on my understanding of economics, human nature, and the skills of people working in tech.”featured in #616
Dumb Leadership Mistakes I’ve Made
- Laura Tacho tl;dr: (1) Dismissing intuition. (2) Data-driven theater. (3) Trying to be smart instead of making other people smart. (4) Not utilizing experts soon enough. (5) Not realizing that I’m not an engineering leader.featured in #615
Developers Don’t Need More Documentation
- Dennis Pilarinos tl;dr: Docs get written, but answers stay hard to find. The problem isn’t the docs themselves. It’s that the context developers need is scattered, outdated, or missing entirely. Why does this keep happening? And what’s the alternative?featured in #615
How To Provide Feedback On Documents
- Will Larson tl;dr: At Carta, the following approach helped reduce resistance to cross-team document sharing.(1) When providing document feedback, prioritize helping the author over other concerns. (2) Skim the document to learn its structure. (3) Leave focused comments with the following structure: the concern, its importance, and why it matters. (4) Limit to 3-4. If it’s over 3-4 comments, schedule a conversation or increase your threshold for a comment.featured in #615
Characterizing A High-Functioning Team
- Abi Noda tl;dr: “In order to understand how individual efforts combine to determine group outcomes like productivity, the authors conducted their research in multiple phases. First, they developed a way to measure collaboration (i.e., who works with who, and when and how they interact). Next, they characterized what it means for a team to be “high functioning.” And finally, they explored the impact of team functioning on other outcomes such as productivity and satisfaction.”featured in #614