Summary of "The Art of Doing Science and Engineering"

Core Idea

Greatness is learnable: Excellence comes from preparation, directed vision, and relentless practice—not innate genius.
Master fundamentals deeply: Surface knowledge blocks progress; understand core principles before specializing or using tools.
Anticipate the future: Study how great minds think and where your field is heading; prepare for what's coming, not what was.

Question accepted rules relentlessly—ask "why?" even when experts say "that's how it's done."
Do back-of-the-envelope calculations to reality-check claims and build intuition.
Saturate yourself with hard problems; let your subconscious work on them constantly.
Luck favors the prepared mind—recognize opportunities only if you've mastered the fundamentals.

Define your vision of future greatness early; a directed path gets you proportional to n, not √n.
Regularly ask: "What's possible? What's likely? What's desirable?" in your field.
Work on important problems only—trivial work is time lost from your career.
Keep 10-20 hard problems simmering; drop everything when a breakthrough appears.

Learn by grasping core principles, not memorizing details.
Teach (and learn from) by studying style and thinking method, not just results.
Master fundamentals by hand before outsourcing to tools—integration by hand teaches pattern recognition.
"The purpose of computing is insight, not numbers"—seek understanding over output volume.

Don't optimize components; optimizing one part often degrades the whole system.
Design for change and unknowable future upgrades; assume specs will be exceeded.
Assume published accuracy is overstated by 5x; pretest all data for inconsistencies.
You get what you measure: Think through unintended consequences before implementing metrics.
Use full dynamic range of rating scales; understand how measurement distorts reality.
Recognize when definitions shift in time-series data—flag this explicitly.

Watch emerging tech early; recognize paradigm shift implications before experts do.
Ask "what else must change?" when adopting new technology systemically.
Verify test equipment is as reliable as what's being tested.
Simulations are essential but dangerous; always ask "Why should anyone believe this?" before trusting results.
Use analogies aggressively—most breakthroughs combine unrelated fields.

Master three presentation modes: formal talks, written reports, informal conversations—selling ideas is essential.
Schedule regular "great thinking" time away from daily tasks to stay strategic, not just tactical.
Build diverse knowledge with multiple retrieval "hooks" for cross-disciplinary insight.
Tolerate ambiguity: believe in your field AND see room for improvement simultaneously.
When promoted, actively avoid blocking the next generation; let go of exclusive knowledge.

Pick one hard problem worth 5+ years of effort; define your vision of solving it.
Audit your learning: Are you grasping principles or memorizing details? Shift to first principles immediately.
Design an experiment to validate a core assumption in your work; use back-of-the-envelope math first.
Block 2-4 hours weekly for strategic thinking away from urgent tasks—let problems percolate.
Identify what you're measuring and unintended consequences; adjust metrics to optimize global performance, not just components.