On Thursday, June 16, 2022, at 10:39 AM, Greg Staskowski wrote: > Ok.. So what are the rulesetz for a "humble machine" > > Don't hurt creators? > Don't run bad code? > Follow all your rules? > > I'm genuinely curious First, to make AGI, one may look to the SOTA AIs like I posted here > https://www.longecity.org/forum/topic/114869-2022-may-20-here-is-where-we-are-with-agi-now/
That Blender one by Facebook has goal/ persona forcing it to say certain domain ex. carssss i like cars i car hobbyist and i keep talking about cars........and it makes it sound very professional and focused BTW. Well, you do just set it the ones you want....ex. human images of nude bodies, food, immortality, and this way it has it on its mind 24/7, as its goal/ thing it wants to about. It does learn much these from a large amount of internet data, actually. But you may be better off putting some hard stalled ones in yourself. Yes, it could find a bad way to get those goals, but so far large AI models seem well grammared and intelligent, even with the biases/misunderstanding part. ------------------------------------------ Artificial General Intelligence List: AGI Permalink: https://agi.topicbox.com/groups/agi/Td8db5de3bbc0c6ae-M43c139043a74f98393044bcc Delivery options: https://agi.topicbox.com/groups/agi/subscription
