I'm Tired of Talking to AI: Three Episodes Where People Forwarded AI Answers

An essay published on the Orchid Files blog on 22 May 2026 makes a simple argument: the author is tired of talking to AI, and even more tired of people who pass AI answers off as their own.

Three episodes

The first involved malicious code. The author found GitHub repositories that were spreading malware and asked an AI tool what to do about it, but the answer was of no use. So he opened a GitHub discussion instead. Someone replied, and the reply was the exact same text the AI had produced. He pointed that out, and the comment was deleted. Then another person replied, again with the same AI answer.

The second episode happened at work. As a developer at a company, he asked the business owner a question about a business task. The owner sent back a screenshot of a ChatGPT answer. The author replied that it had nothing to do with his question and that everything in it was wrong. A minute later the owner sent another ChatGPT screenshot. He had not read the AI answer at all; he simply took a screenshot and forwarded it.

The third episode was a conversation on Reddit about one of the author's posts. He replied, the other person wrote again, he replied again, and after a few messages he realised he was talking to an AI agent.

What the pattern shows

The common thread is not that the tools are useless. It is that the answer arrives before anyone has understood the question. In the GitHub case the generated text displaced a real discussion of a security problem; at work a screenshot replaced reading; on Reddit a person was not there at all. Each time, the human step that gives an answer meaning, checking it against the problem, is skipped.

The author's conclusion

He writes that he wants to talk to real people. But even when he does, his questions are forwarded to AI and the AI answer comes back. The essay ends there, without advice or a fix, and that is part of its point: the friction of a real exchange is what is being lost.