Quote:
Originally Posted by Ekco
[You must be logged in to view images. Log in or Register.]
yeah, most of the bots on twitter pretending to be badass, don't post autonomously its QA'd and approved by a person before it's tweeted out
cause it will do stupid shit and misinterpret things, or like mine would respond to someone with text, then respond to the tweet it made and put the hashtags in that tweet for the previous response
i don't know how i even fuckin fucked that up like that but i did, so shit still needs humans curating it and it's mostly tech demos and fancy trailers with "release date announced soon" like any half finished game presented at E3 or something
|
I think to manage it you need to try a daisy chain of AI.
Like,
1. Ai that manages the goals/directives of sub AI
2. Sub AI to check reddit every now and again.
3. Sub AI to look for the right types of posts to interact with and create's a ledger.
4. Sub AI that checks the ledger for posts to respond to.
5. Sub AI that writes the responses.
Like basically ai that manages 1 small aspect, of ai, that does the same until you get to the granular stuff.
Think of these sub AI's as organs that together make a whole but alone do 1 thing only.
I think that's how to make a sentient AI. But I haven't seen it done, would be interested in experimenting with you to see if we can make a person. I have it all mapped out!