Dimebag wrote: ↑Fri Jan 06, 2023 11:38 pm
I have noticed the occasional interaction, but his posts are not always relevant.
If someone was taking the Turing test and trying to determine whether the responder was human or program, they might come to the conclusion that dattaswami was a program for that reason.
The truth is, we don’t know. Many of the posts we see here are also word for word copy pastes of some of his blogs from his website.
The intention of dattaswami does not appear to be conversation, which is what is required to pass a Turing test.
He seems to have no intention to interact, only to post his blogs here.
I've actually found him to be more responsive than Peter Kropotkin, for example, but the bulk of his participation is, yes, copy paste. Further his reponses don't really fit the criticisms, questions, arguments of the people he is supposedly responding to. Responses to his posts tend to trigger short or long lectures. Which is a perfect pattern for a
not yet passing the TT AI. I think this is precisely how a not successful program would respond. It's related. It's not random. But the program is not able to tease out the exact criticism. It doesn't quite know how to respond to the specific point made. So, it pulls 'true' stuff out of its memory and throws it down as if it is a response.
Now do I really think DS is an AI? Well, I would guess he's not. But I realized that I think an AI, with access to the swami's blogs, could probably do as well, today, as DS does in responding here.. IOW his abilities as a participant in a
dialogue are not out of reach of current AIs.