Don't paste the AI, please
Generative AI is increasingly being used as a shortcut in workplace and online communications, with people pasting unedited chatbot responses into emails, chats, and code reviews. Many commenters argue this erodes trust, shifts the burden of comprehension and validation onto others, and undermines the value of human judgment, even as they acknowledge legitimate uses like drafting, translation, or adding context. The thread also highlights the irony that anti–AI-slop campaigns themselves may use AI, and calls for clearer social and organizational norms: label AI use, don’t forward unvetted output, and prioritize concise, human-authored summaries over raw model text.
Perceived irony and AI “slop” accusations
- Many commenters say the site itself “reads like AI,” especially the “angry” version, citing tone, structure, and metaphors as typical LLM output.
- AI detectors (e.g., Pangram, ZeroGPT) are mentioned, with disagreement over their reliability and the fear of false positives.
- Some find it deeply ironic or hypocritical that an anti–AI-paste site appears partially AI-written; others argue what matters is whether the content is useful and honest about its creation.
Core complaint: being used as a “meat proxy”
- Strong agreement with the core critique: blindly pasting AI output into chats, email, or PRs offloads thinking, validation, and summarization onto the recipient.
- This behavior is likened to LMGTFY, “slop-bombing,” and “meat proxy” work where a human just relays model outputs without understanding them.
- Several note that people ask humans specifically for their judgment, context, and concise synthesis, not generic model text.
Defenses of pasting AI and counterpoints
- Some argue that with rich personal context, tuned agents, or internal data, their AI answers are not generic and can add value others can’t easily reproduce.
- Others see verbatim AI replies as appropriate pushback against lazy, obvious, or bad-faith questions, analogous to “Google it yourself.”
- Critics respond that this still hides how little effort or understanding the sender has, erodes trust, and often creates long, wordy, low–signal messages.
Workplace norms, policies, and etiquette
- Multiple commenters describe internal “AI principles”: write in your own voice, own and fully review outputs, don’t slop-bomb, and be the expert rather than delegating judgment.
- There’s frustration with colleagues, managers, and offshore devs who paste unvetted AI answers in Slack, docs, code reviews, and customer comms.
- Others report the opposite: LLMs help poor communicators add missing context, which can make support and ops work easier.
Broader concerns: thinking, literacy, and culture
- Some fear AI dependence weakens writing, thinking, and learning (“atrophy of the thinking muscle,” anchoring effect, “pseudomentalizing”).
- Others see AI as just another tool (like calculators or spellcheck) and blame organizational culture, not tools, for bad communication.
- Several meta-links and alternative “no-slop” sites are shared, reflecting an emerging mini-genre of anti–AI-slop etiquette pages.