←back to thread

If you have a Claude account, they're going to train on your data moving forward

(old.reddit.com)
439 points diggan | 3 comments | 29 Aug 25 11:39 UTC | HN request time: 0.001s | source
Show context
ratg13 ◴[29 Aug 25 12:44 UTC] No.45063367[source]▶
I can understand training AIs on books, and even internet forums, but I can't help but think that training an AI on lots of dumb questions with probably an excessive amount of grammar and spelling errors will somehow make it smarter.
replies(4): >>45063503 #>>45063715 #>>45064070 #>>45068370 #
1. nrclark ◴[29 Aug 25 13:15 UTC] No.45063715[source]▶
Depends on how you’re using the data. There’s a pretty strong correctness signal in the user behavior.

Did they rephrase the question? Probably the first answer was wrong. Did the session end? Good chance the answer was acceptable. Did they ask follow-ups? What kind? Etc.

replies(2): >>45064137 #>>45064409 #
2. dudefeliciano ◴[29 Aug 25 13:51 UTC] No.45064137[source]▶
> Did the session end? Good chance the answer was acceptable.

Or that the user just ragequit

3. vb-8448 ◴[29 Aug 25 14:13 UTC] No.45064409[source]▶
I'm used to doing the same task 4 or 5 times (different sessions, similar prompts), and most of the time the result is useless or completely wrong. Sometimes I go back and pick the first result, other time none of them, other time a mix of them. I'm wondering how can they extract value from this.