Wrong, not broken

Lobsters·5 points·sp6370·4 days ago·aws.amazon.com

Software operations has always asked one question: Is it broken? AI agents change that. Here’s the shift toward measuring correctness at the level of the run, and the thinking behind Amazon CloudWatch Omni.

Read the full article at aws.amazon.com →

#broken #wrong #vibecoding #testing #software #always #asked #operations

Related stories

Lobsters Interview with Susam
Lobsters · 86 points · about 1 year ago
Wrong, Not Brokenaws.amazon.com
Hacker News · 1 points · 4 days ago
Ask HN: Will software still matter when AI write all our programs?
Hacker News · 1 points · 7 days ago
Show HN: Lathoa, a math app for kids where the AI is wrong on purposelathoa.ai
Hacker News · 28 points · about 12 hours ago
Show HN: Local Software Factory – running multiple coding agents in parallelgithub.com
Hacker News · 1 points · 7 days ago
Why the Best Software Advice Is the Hardest to Followdev.to
DEV Community · 27 points · 9 days ago

Related topics

ai2032 always29 agents430 promises7 coding agents34 agent432 asked30 operations9 coding195 software195 wrong57 shipping20