Future Timeline

We both posted at the same time!

wjfox wrote: ↑Wed Jan 11, 2023 10:26 pm We both posted at the same time!

deleted

"Exclusive: OpenAI Used Kenyan Workers on Less Than $2 Per Hour to Make ChatGPT Less Toxic"

https://time.com/6247678/openai-chatgpt ... reddit.com

These are the people deciding morals of AI

Personally there is no point to me becoming a luddite, AI will come before any kind of cultural change I want. "The end is near".tm will we overtime get Heaven or Hell that is the question.

wjfox wrote: ↑Thu Jan 26, 2023 7:39 am

I wonder what the false positive rate will be for this kind of "ChatGPT detection" algorithm. It may be able to consistently identify responses made by ChatGPT, but how often will it pick out someone's own writing and identify that as ChatGPT output? Especially in the kind of sterile and couched language I see in my work email circles, a lot of what I see looks similar to how ChatGPT writes (or rather, the other way around). For people who naturally write in a similar style, would they get improperly flagged?

Though I'll admit I don't know much about these programs or the in's and out's of how this detection system will work so it might not be a legitimate concern at all.

[Edit] It looks like the watermark will be on the output side, built into ChatGPT, so my concern doesn't apply to the quoted tweet. Instead, I should direct my concern to the detection bots like the ones mentioned elsewhere like a "plagiarism detector".