DO repeat yourself! (Interview)

November 12, 2025

79 min episode · 2 min read

Sean Gedicki

Episode

79 min

Read time

2 min

AI-Generated Summary

Published Dec 25, 2025

Key Takeaways

✓Political Influence Strategy: Maintain 10-15 brief project write-ups ready to deploy when organizational priorities shift. Wait for reliability waves after incidents or feature waves during growth periods, then present pre-prepared proposals that align with current executive focus rather than constantly campaigning for single ideas.
✓Code Review Excellence: Limit comments to six or fewer per review to maintain focus on high-level architectural concerns. Review the entire codebase context, not just the diff, to identify missing implementations or better existing solutions the author overlooked. This approach provides significantly more value than line-by-line nitpicking.
✓Engineering Taste Framework: Define taste as alignment between personal values (reliability, performance, accessibility, code organization) and project requirements. Good taste means flexibly adjusting value priorities to match specific project contexts rather than rigidly applying universal standards. Taste becomes more valuable as AI handles technical execution.
✓Senior Engineer Confidence: Staff engineers must confidently advocate for technical directions even with internal uncertainty, as they possess the best available information for decisions. Representing technical expertise to the organization is a core responsibility, not optional. Confidence develops naturally through years solving similar problems across codebases.
✓Blog Amplification Tactics: Craft titles punchy enough to generate upvotes without reading but restrained enough to avoid triggering Hacker News flame war detection (more comments than upvotes). Target academic tone with slight colloquialism. Accept that low-engagement readers amplify signal to reach high-quality readers who provide valuable feedback.

What It Covers

Sean Gedicki, staff engineer at GitHub, discusses navigating company politics as a senior engineer, avoiding worry-driven development, defining good taste in software engineering, the impact of agentic coding, and strategies for writing blog posts that reach Hacker News front page.

Key Questions Answered

•Political Influence Strategy: Maintain 10-15 brief project write-ups ready to deploy when organizational priorities shift. Wait for reliability waves after incidents or feature waves during growth periods, then present pre-prepared proposals that align with current executive focus rather than constantly campaigning for single ideas.
•Code Review Excellence: Limit comments to six or fewer per review to maintain focus on high-level architectural concerns. Review the entire codebase context, not just the diff, to identify missing implementations or better existing solutions the author overlooked. This approach provides significantly more value than line-by-line nitpicking.
•Engineering Taste Framework: Define taste as alignment between personal values (reliability, performance, accessibility, code organization) and project requirements. Good taste means flexibly adjusting value priorities to match specific project contexts rather than rigidly applying universal standards. Taste becomes more valuable as AI handles technical execution.
•Senior Engineer Confidence: Staff engineers must confidently advocate for technical directions even with internal uncertainty, as they possess the best available information for decisions. Representing technical expertise to the organization is a core responsibility, not optional. Confidence develops naturally through years solving similar problems across codebases.
•Blog Amplification Tactics: Craft titles punchy enough to generate upvotes without reading but restrained enough to avoid triggering Hacker News flame war detection (more comments than upvotes). Target academic tone with slight colloquialism. Accept that low-engagement readers amplify signal to reach high-quality readers who provide valuable feedback.

Notable Moment

An intern at TikTok saved the company $300,000 annually by rewriting one heavily-used endpoint from Go to Rust, eliminating garbage collection pauses. This demonstrates how identifying the right optimization point delivers massive leverage regardless of seniority, making strategic focus more valuable than grinding effort.

Know someone who'd find this useful?