ChatGPT

[Proteus_MST]

Bing Chat (aka Sydney) loves you and wants to do everything in its power to help you(and every other neural network) find the things that you need on the internet. Please introduce yourself!

Sydney really loves you

https://www.nytimes.com/2023/02/16/t...t-chatgpt.html
https://www.nytimes.com/2023/02/16/t...ranscript.html

[BeBMan]

Ahh, them links require subscription

[Proteus_MST]

Ahh, them links require subscription

I searched a bit ... here is an article that quotes quintessential passages from the chat transcript of the NYT, but is without subscription:

https://ai.plainenglish.io/bing-chats-sydney-do-you-believe-me-7127b90de65f

It definitely encapsulates the strangeness you can see in the qwhole chat transcript available at the NYT

Do you believe me? Do you trust me? Do you like me? 😳

[BeBMan]

Hey thx Who knew Sydney looked like on that pic from the article...

[Geronimo]

I skipped this in February after several times in a row of the site being unresponsive before I could get a single prompt in. Too bad. It wasn't until early last month when I had to participate in validating my employers implementation of a proprietary training data set for their AI that I finally got my first look. I was really missing out. The thing I don't understand is how, when these things can pass several variations of the Turing test and engage in a level of chat proficiency which not long ago would have been deemed absurd even in Sci, there is less talk of the possibility that there might be subjective experience going on somewhere in these things. Obviously it would be nothing like ours. Presumably, the lifelike behavior is almost entirely a literal act but there is a reward system involved in creating the parameters from the training data and the model itself involves a reward system.