I just watched a simple website called dontquotetheai.com climb to the top of Hacker News with 161 points and 101 heated comments. The message? Stop dumping raw LLM output on other humans.
This isn't some anti-AI crusade. It's something more interesting: the developer community is establishing etiquette for the AI age.
I've been guilty of this myself. Last week, I pasted a 400-word ChatGPT response into our team Slack when someone asked about database indexing. The response was technically correct but felt... hollow. Like I'd outsourced my brain to a statistical parrot.
<> The irritation is not just quality; it's the feeling that the sender has outsourced the conversation rather than participated in it./>
That line from the research nails it perfectly. We're not just sharing information anymore—we're transferring cognitive load. When you paste raw AI text, you're making the reader do the work of:
- Figuring out what's actually relevant
- Verifying if it's correct
- Determining your actual opinion
- Parsing through generic fluff
The technical implications are fascinating. Product teams building AI-assisted tools need to rethink their UX patterns. Instead of big "Generate" buttons that dump walls of text, we need:
1. Editing-first workflows that encourage synthesis
2. Provenance indicators showing AI vs human contributions
3. Confidence scoring for uncertain claims
4. Summarization tools that extract key points
The business angle is equally compelling. Companies deploying AI into customer support are walking a tightrope between efficiency and authenticity. A poorly handled AI reply doesn't just waste time—it damages trust.
I'm seeing this play out in real-time on engineering forums. Stack Overflow conversations increasingly feel like AI echo chambers. Someone asks a specific question about React hooks, gets a 500-word ChatGPT response covering every possible hook use case. None of it addresses the actual problem.
The controversy brewing is subtle but important. This isn't about banning AI—it's about establishing social norms for a world where anyone can generate authoritative-sounding text instantly.
Some pushback arguments I expect:
- "But the AI answer is more complete than what I could write"
- "People should just learn to skim better"
- "This is just productivity optimization"
Missing the point entirely. Communication isn't about information transfer—it's about human connection and shared understanding.
The market opportunity here is massive. We need tools that help humans collaborate with AI rather than just delegate to it. Think AI that helps you write better responses, not AI that writes responses for you.
My Bet: Within 18 months, major platforms (Discord, Slack, GitHub) will add native "AI-assisted" labels and editing workflows. The one-click "paste AI response" era is ending before it really began. Developers are too opinionated to accept slop, even efficient slop.
