alCannium27
Reaction score
299

Profile posts Latest activity Postings Media Awards About Agora Gang Tags

  • So the Poles got the world's first AI robot CEO. Nothing more than a marketing scheme for now. But with Sam Altman looking to set up shop over there to push for more AI integration into business operations, it won't be long before all businesses are automated. How long before humans are made obsolete? A century? A decade?
    mika-ron-dictador-2.jpg.webp
    Taleisin
    Taleisin
    @MySpace Tom I disagree with the idea that we cannot define intelligence. There are many alternate definitions used in different fields, but each is valid for that purpose. AGI need not be conscious (though I think it's likely it would be immediately, no matter the mechanisms involved), all it needs is to be an autonomous agent that can learn and generalise across modalities and domains. AGI is not "making a human robot", it's just making something that can act on its own, has representational self-awareness, and is smart enough to not fail when presented with new information.
    Taleisin
    Taleisin
    @MySpace Tom here's some info that is informing my thinking on this



    View: https://www.youtube.com/watch?v=jLiHLDrOTW8

    MySpace Tom
    MySpace Tom
    @Taleisin (Re: the 1st comment; will check the links later)

    A problem is that the agent needs a motive to keep developing. Human motives are tied to instinctual needs for survival and reproduction (granted, with a lot of abstraction in the middle). This implies that there is the undesirable outcome of mortality. A machine would need some kinds of instinct programmed into it first in order to be functional. Plus, to comprehend its environment, it would need a body that has different senses that are interpreted by the instinct on some level (fire = painful, avoid). I don't think we can build such a machine

    I think that intelligence in the AGI sense does need consciousness in order to truly operate (I think, therefore I am, after all). Otherwise, we would just have a machine that has Pavlovian responses to different stimuli which, while still useful, doesn't necessarily mean that it is an intelligent machine. This, of course, leads us to the inescapable problem of The Chinese Room
  • Loading…
  • Loading…
  • Loading…
  • Loading…
  • Loading…
  • Loading…