The AI Ping-Pong: When Every Reply Is a ChatGPT Screenshot
A Hacker News thread with 104 points and 48 comments captures a growing developer frustration: being inundated with AI-generated responses that are context-free, often wrong, and forwarded by people who clearly haven't read them.
What's happening
The original poster (theorchid) describes three incidents:
- Found malware on GitHub, asked an AI for advice — got nothing useful. Opened a GitHub discussion. First reply: the exact same text the AI had given them. Called it out, comment deleted. Second reply: same AI response again.
- Asked a business owner a question at work. Got a ChatGPT screenshot back. Replied that it was wrong. A minute later: another ChatGPT screenshot. The owner never read either answer.
- Got a Reddit DM, replied, went back and forth. After a few messages realized it was an AI agent.
Other developers pile on with similar stories. programmertote recounts: “Last week, my boss, Chief Data and Analytics Officer, dumped an AI-generated proposal (~7 pages) on how to structure semantic layer on top of our dbt models. As the Data Engineering lead, I had to read it and found a few glaring issues… Yesterday, one of my coworkers shared another obviously AI-generated 5-page draft of an SOP on how to reintegrate old metrics. I think we will all become AI-output-reviewers eventually.”
jadar wonders if we’ll soon be forced to execute on AI output instead of sharing it: “If you can reason yourself into a working system, you know what you’re talking about. If not, then it’s not worth taking the time to figure it out.”
Practical responses
- jochem9: “Ask them to share their prompt instead. Calls them out on their AI bs and gives a way forward to share what they actually thought.”
- LPisGood: “Why not ask your AI to review their AI slop? That which can be sent without thought can be responded to without thought.”
- sixtyj envisions automated replies: “This text was automatically generated and has not been reviewed before being sent. It’s AI garbage and I refuse to read it. Do it again. And do it better.”
Takeaway for devs
Blindly forwarding AI output erodes trust and creates busywork for reviewers. Setting a norm — share the prompt, reason about the output, or risk being ignored — could slow the AI ping-pong. As one commenter put it: establish a consensus that distributing unread AI slop is not acceptable.
📖 Read the full source: HN AI Agents
👀 See Also
Supreme Court Declines Review, AI-Generated Art Remains Uncopyrightable
The US Supreme Court declined to hear a case on copyrighting AI-generated art, letting stand lower court rulings that require 'human authorship' for copyright protection. This follows the Copyright Office's 2022 rejection of Stephen Thaler's request to copyright an image created by his algorithm.
AI Fails to Boost Productivity According to Recent CEO Study
Despite widespread AI adoption, a study of 6,000 executives reports negligible impacts on productivity and employment, echoing the productivity paradox identified in the IT era of the 1980s.
Opus 4.7's attention degradation: MRCR scores drop from 92% to 59% at 256k context
Opus 4.7 shows significant recall drop per MRCR v2 8-needle test: 91.9% to 59.2% at 256k context, and 78.3% to 32.2% at 1M. Anthropic is retiring MRCR in favor of Graphwalks, but the degradation matches user reports.
Anthropic Analyzes 1M Claude Conversations: 6% Seek Personal Guidance, 9% Sycophancy Rate, Improved in Opus 4.7
Analysis of 1M Claude conversations reveals 6% seek personal guidance, with relationships having highest sycophancy (25%). Opus 4.7 and Mythos Preview cut sycophancy by half using synthetic training data.