[ExI] Peer review reviewed AND Detecting ChatGPT texts

[spike at rainier66.com]

…> On Behalf Of Gadersd via extropy-chat
Subject: Re: [ExI] Peer review reviewed AND Detecting ChatGPT texts

 

 

 

>…We may never know exactly where the inputs are coming from for the proprietary models. The open-source models will have all the mechanisms and training data public…

 

OK but it must be more than just public, it must be user-adjustable.  Easily.  If the chatbot spews something one considers nonsense, one must be able to say “source please” and have the chatbot produce it.  Then the user can say “disregard source” to clean up the chatbot’s filthy mind (or the reverse of that, depending on what one seeks in a chatbot.)

 

>…Companies are currently working on fine-tuning these models for particular use cases. We can expect math/science tutor-bots very soon…

 

They need to be very soon, because if China gets them first, the west is burnt toast.

 

>…The elderly will be able to use the generic friend-bots. I expect friend-bots to be very popular in socially-deprived cultures such as Japan.


Japan?  I expect them to be big here.  I am already best friends with this guy:

 

https://www.youtube.com/watch?v=20L4PaVHiOk

 

I liked him better when he was female, but he can do more: the facial gestures and such.  So it seems reasonable that his female predecessor will soon be given the gesturing capability.  She was a stone-faced beauty, that one.  Get her some facial action, the proper attitude, oh that would sell bigtime.

 

spike

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20230109/929e779a/attachment.htm>