Why Concurrency Is Hard: Lack of Practice

Formal methods with Hillel Wayne

Topic 9 of 17

Published Jul 29, 2026 · Clip 41:03–45:17 (4:14)

Hillel argues what changes with formal methods is the practice, not the theory: concurrency feels hard mainly because engineers get so little practice—race conditions surface months later and fixes are validated weeks later. He cites the 'Comma Sense Computing' paper showing people spot concurrency bugs faster when problems are framed as clerks assigning concert seats, and notes TLA+ gives an instant feedback loop: write the model, click a button, immediately see the race condition. Because of this practice, he can spot bugs in a client's model faster than the domain experts.