• Public

  • Public
  • Network
  • Groups
  • Featured
  • Popular
  • People

Screenshot from a Reddit post showing the following text (warning: long text with several paragraphs): "AI Slop PR's are burning me and my team out hard, anyone else experiencing this? Background: Current role is a TL (dev/manager hybrid at this place), my team has a large amount of domain ownership so we are constantly pinged for PR reviews. Lately there has been a huge push for teams to adopt tools like Cursor, the problem is that while yes they can generate code, it is just lately rapidly becoming an endless stream of AI slop. In the last few weeks: Multiple 5k+ line PR's that should be sub 100 lines PR's that have tons of changed files that in some vibe coding iteration were dropped or my new favourite thing endless redirection where multiple things don't actually do anything. Very scary PR's where the AI did something extremely dangerous i am assuming to make tests work or something. For example one of the PR's actually did such a very subtle change where it aborted early in a middleware basically skipping most of AuthZ, then mocked out a good chunk of the AuthZ in tests which caused tests to pass. AI hallucinating external services, then mocking out the hallucinated external services. Forcing me to go look up other repos/service maps and validate that yes this api endpoint actually exists. AI's ignoring project architecture and structure, dumping files everywhere, or ignoring coding styles."