PulseAugur
LIVE 15:26:22
meme · [1 source] ·
0
meme

AI slop: Prompt engineer feeds critique back into LLM, generating worse code

A user on Mastodon expressed frustration with the outcome of critiquing LLM code. They described a scenario where a "prompt engineer" used the original code and the critique as input for an LLM, resulting in over 4000 lines of code that were perceived as worse than the original. The user characterized this experience as part of a "hellish reality" associated with AI. AI

Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →

RANK_REASON The content is a personal anecdote and opinion shared on a social media platform, lacking broader newsworthiness or verifiable facts.

Read on Mastodon — mastodon.social →

COVERAGE [1]

  1. Mastodon — mastodon.social TIER_1 · [email protected] ·

    If you could have predicted that the outcome of critiquing an LLM's code will be a "prompt engineer" feeding the original code together with my critique into an

    If you could have predicted that the outcome of critiquing an LLM's code will be a "prompt engineer" feeding the original code together with my critique into an LLM and coming back with over 4000 lines of even worse code, you're more experienced in this new hellish reality than I…