# Anyone have similar issues to this — retained record

> Retained reddit record from reddit:ChatGPT.

Canonical: https://highsignal.app/data/records/a7b148479390bbb2

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:ChatGPT)  published 2026-10-31 15:11 ingested 2026-11-01 02:36

[original source ↗](https://www.reddit.com/r/ChatGPT/comments/1xkcppr/anyone_have_similar_issues_to_this/)

## retained text



```
The stuff where it'll say something then go and say that itself is wrong later. Like I'll see sometimes, and I don't have a screenshot but it could be similar to this

  
Prompt: Name five English words beginning with a.

  
Answer: Apple, Athlete, Art, Arm, Dog (Oh wait dog doesn't actually start with an A!, I meant Ambulance)​

  
Does anyone else have this problem? Wouldn't it be easy to have it just read through everything it says before giving it to the user? Because how it is now i think it'll make a mistake but because it is likr procedurally generated it can't change it but idk 🤷‍♂️

  
I just find it annoying because like in the screenshot it'll give you an answer but if you don't read until it corrects itself then you'll get an incorrect answer.
```



document fetched 2026-11-01 00:20 · canonical https://reddit.com/r/ChatGPT/comments/1xkcppr/anyone_have_similar_issues_to_this
