← / writing --tooling

vibecheck: is claude code actually faster than typing?

a week of timing every commit. the answer is: yes, but only for a specific kind of work — and the gap is smaller than the discourse implies.

i spent a week timing every commit on a side project. here’s what i actually measured.

the setup

  • one engineer (me), one codebase, ~12,000 lines of typescript
  • track time-to-first-commit for each task, split by tool: hand-typed vs claude code
  • 23 tasks total over 5 days, average task size: ~40 lines changed

the headline number

task typehand-typed (avg)claude code (avg)delta
boilerplate / CRUD14 min4 min-71%
algorithmic22 min18 min-18%
ui polish9 min11 min+22%
debugging28 min35 min+25%

claude code wins on CRUD and loses on debugging. neither result will surprise anyone who’s used it for a week. but the magnitude is worth noting: the discourse undersells the CRUD speedup and oversells the debugging speedup.

the catch

i’m not counting review time. if i included the time to read claude’s diff carefully — which i do, every time — the CRUD delta narrows from -71% to closer to -50%. still a clear win, just less dramatic.

the verdict

vibecheck rating: 8/10. clear win on routine work, neutral-to-negative on hard work, and the ergonomics are good enough that i don’t switch back even for the tasks where it’s slower. the loss aversion of context-switching tools is real.