~/talk

Why Humans Can't Keep Up With AI-Written Code

links:none on file
slides:none on file

Agents produce 2,450% more lines of code but releases rise only ~30%, and five failure modes matter: humans agree with wrong AI 4 out of 5 times, agents fail on very large codebases (accuracy collapses past ~50k LOC), errors compound over long task chains, LLM judges flip answers 7–34% of the time, and exploit windows are near zero. Sonar’s approach uses an intent+guide rails+verification+evidence formula and products like SonarVortex and SonarQube to cut token costs 36% and commit issues 92%.

AI-written summary · about

  1. 0:00Intro
  2. 0:26The ball bearing that changed car factories
  3. 2:23Software already borrowed these ideas
  4. 2:55Agents break the factory
  5. 5:19Problem 1: Human in the loop doesn't scale
  6. 6:45Problem 2: Agents struggle with big codebases
  7. 8:04Problem 3: Errors compound
  8. 9:06Problem 4: LLM judges are unreliable
  9. 10:03Problem 5: Time to exploit is near zero
  10. 10:25The code confidence formula
  11. 12:36Results from customers
  12. 15:49Code is abundant, confidence is scarce
~/search
type to search 6,518 conferences, 143,918 speakers and 192,133 talks
~/help
cmdctrl k
search every conference, speaker and series
/
focus the filter search
h
home
c
conferences
s
speakers
e
series
t
light / dark — the status bar button does the same
← →
on a talk: 5 seconds back / forward
0–9
on a talk: jump to 0%–90% of the recording
?
this help
esc
close this, then clear every filter — or just the search, while you are typing in it

Ignored while you are typing. Pages without a filter rail have no search to focus.