You Are the CPU: What CS Can Teach Us About Ourselves to Improve Developer Efficiency
summary
Shows that developers now outvalue hardware—1969 System/360 cost $5M (~$40M today) with 64 KB RAM versus an $800 iPhone 13—so optimize human time not just CPU. Cites Accelerate metrics: deployment frequency, lead time <1 hour and <15% error rate with MTTR <1 hour. Maps Big O to toil, L1/L3 cache to short‑term memory, and IO cost (RAM ≈6 minutes, a cross‑country packet ≈4 years on human scale). Recommends reduce context switching (70% waste across five projects) and ship smaller batches.
AI-written summary · about
chapters (11)
- 0:00Introduction to CPU analogy
- 1:54Developer as the expensive resource
- 3:40Measuring team performance metrics
- 4:42Big O and developer efficiency
- 6:33Eliminating toil in workflows
- 9:03CPU caches and cognitive load
- 11:34Synchronous vs asynchronous work
- 13:40Reducing context switching
- 15:28Cost of I/O in human time
- 17:40Reducing unnecessary communication
- 18:23Ship smaller batches frequently
related talks
51mThe astonishing value of speed
29mEconomies of Speed
46mLet Them Eat Code
10mAchieving engineering velocity by optimising developer experience
44mSoftware Management Lessons from the 1960s
45mThe astonishing value of speed
54mReadable Code: The Secret to Clean, Productive Software
30mOptimizing for Time: Dark Matter