One annoying thing about "AI" slop code being a thing now is that you don't know anymore who to blame for issues in new programs/apps:Was it a human who accidentally left a bug in?Was it generated by an LLM and just never fully tested?Is there maybe some setup issue?At least in the past you had the guarantee that the author(s) had thought it to fully work before going through the whole process of packaging it up and being proud enough to publically release it somewhere, so the first action was to check the setup when things went wrong. But with a ton of people releasing crappy LLM-code before even testing it this is now turned upside down.This shit is the reason why I try to avoid git repos with hints of chatbot usage now... (https://slopscan.ava.pet ftw!)#ai #noai #aislop #opensource #slopscan
Related
“Been seeing a lot more #outages amongst the big tech companies in the last year. I have suspicions this is cause of mor...
“Been seeing a lot more #outages amongst the big tech companies in the last year. I have suspicions this is cause of more bad and insecure #AIgenerated #code being pushed into code...
Don't let the same LLM write and review your code.→ Claude analyzes your last 1,000 PRs to build a team style guide, the...
Don't let the same LLM write and review your code.→ Claude analyzes your last 1,000 PRs to build a team style guide, then writes the code. → Gemini takes over code review and enfor...
Tools like Clipto now allow ChatGPT and Claude to search years of meeting recordings, changing how engineering teams lev...
Tools like Clipto now allow ChatGPT and Claude to search years of meeting recordings, changing how engineering teams leverage institutional knowledge. The core trade-off centers on...