#25 | Sunday reads for EMs

My favourite reads of the week to make your Sunday a little more inspiring.

Dec 07, 2025

👋 Hey, it’s Stephane. This is a new series in which every Sunday I share with you my favourite reads of the week. To accelerate your growth see: 50 Notion Templates | The EM’s Field Guide | CodeCrafters | Get Hired as an EM | 1:1 Coaching

Paid subscribers get 50 Notion Templates, The EM’s Field Guide, and access to the complete archive. Subscribe now.

The Math of Why You Can’t Focus at Work (Can Duruk)

tl;dr: Your productivity is determined by three parameters: λ (interruptions/hour), Δ (recovery time), and θ (minimum focus block needed for meaningful work). At λ=2 and Δ=20min, you might only get one 60+ minute focus block in an 8-hour day. Most “productivity problems” are actually environment design problems.

What Actually Makes You Senior (Matheus Lima)

tl;dr: Senior engineers reduce ambiguity, everything else is downstream. Mid-levels crush well-defined problems; seniors take “we should probably think about scaling” and turn it into two small projects. The invisible work of derisking happens upfront through questions like “what are we assuming that might be wrong?”.

Writing a Good CLAUDE.md (Kyle Mistele)

tl;dr: Claude often ignores your CLAUDE.md if it’s bloated with non-universal instructions. Research suggests LLMs reliably follow ~150-200 instructions max, and Claude Code’s system prompt already uses ~50 of those. Keep it under 300 lines (they use <60), use progressive disclosure (point to separate docs), and never use it as a linter. The “less is more” principle applies to context engineering too.

21 Lessons from 14 Years at Google (Addy Osmani)

tl;dr: Best non-obvious insight: “If you win every debate, you’re probably accumulating silent resistance” - people stop arguing not because you’re right but because they’ve given up, and they’ll express disagreement in execution instead. Also: treat every metric request with a pair (speed + quality/risk), make glue work visible as bounded artifacts rather than personality traits, and remember that most “slow” teams are actually misaligned teams.

Tech Predictions for 2026 and Beyond (Werner Vogels)

tl;dr: Werner Vogels’ annual predictions is more strategic vision than tactical. The “renaissance developer” framing argues AI won’t replace developers but will raise abstraction levels (like compilers and cloud did), creating more demand not less. Interesting sections on companion robots for elder care and AI tutoring hitting scale (Khan Academy’s Khanmigo reached 1.4M students in year one).