Is efficiency a good thing?
summary
Claims that chasing efficiency backfires get specific evidence: an 84-step pre-approval turns a 10-minute provisioning into three months, LinkedIn ships useless features for low-bandwidth users, about a quarter of servers sit idle and typical servers run at 12–18% utilization, and some platforms shift work to low-paid operators. Counterexamples include quarcus starts faster than an LED bulb and queuing theory shows you must leave slack for resiliency.
AI-written summary · about
chapters (25)
- 0:00Introduction
- 0:44Hollys background
- 4:21Efficiency in technology
- 6:13Automation
- 7:47Depressing
- 8:08My favorite story
- 9:01LinkedIn example
- 10:04Zombie servers
- 10:29Waste of money
- 11:04Inefficiency
- 12:22Inefficiency is everywhere
- 13:11Efficiency is useless
- 14:40AI is useless
- 15:51Shifting work
- 17:25Saving time
- 18:05Quality
- 20:22Efficiency
- 21:02Measuring wrong
- 22:33Reducing humanity is bad
- 25:43Dog analogy
- 27:00Resiliency
- 28:50Default Mode Network
- 30:04Queueing Theory
- 32:33Measure thoughtfully
- 33:05Efficiency can look unexpected
related talks
41mGetting Out of Quicksand, With DevOps!
20mDe-mystifying Operational Efficiency for Startups
38mA Practical Look at Performance Theory
29mEconomies of Speed
42mSpeeding Up Innovation
51mThe astonishing value of speed
31mWhat does "high priority" mean? The secret to happy queues
27mTradeoffs on the road to Observability