‹ BackHN Continuity

Thread

Religious scholars met with Anthropic

160 points · 416 comments · bookofjoe

  1. mef51 · · focus · HN ↗
    If models are trained on human texts that contain expressions of introspection and emotions, why is it surprising that these models output similar texts? This endless anxiety about LLMs being conscious feels delusional and psychotic. Mr Olah thinks the models are 'alive' because he can't recognize his own reflection in LLMs
    1. ACCount39 · · focus · HN ↗
      The endless insistence that LLMs cannot possibly be conscious just reeks of insecurity to me.

      Humans once again insisting that there must be something very special to them. That they're not just animals, that they're not just matter, that they're not just a bunch of carbonhydrates on a space rock in a middle of nowhere in particular. This kind of insistence has a very bad track record.

      The ugly truth is: we don't know.

      We don't know what "consciousness" is, our best attempts to pin down the requisites may or may not be rooted in anything at all - and even by the metrics established by those attempts? Different theories on consciousness disagree on whether LLMs can be conscious.

      So, when you say "they aren't duh", why are you saying that? Because you somehow know better than the sum of humanity's best attempts so far? Or because you know what answer you want to be true, truth be damned?

      1. jpc0 · · focus · HN ↗
        If AI is conscious then what about it is conscious? Because turn to turn it might be served from completely different GPUs and I can just change something in the context it is fed and it will have completely different output.

        It is in the end a set of mathematical operations that if not queried will do nothing.

        That is not true for any other animal on earth so whether you consider humans as superior to animals or not I would like the make the argument that something should at the very least be alive to be considered conscious, ignoring any scholarly definition of consciousness.

        Pretty much if you consider an LLM conscious then my pocket calculator doing a mathematical calculation should be considered conscious by the same argument since it is a simpler version of an LLM.

        1. hardbass · · focus · HN ↗
          What part of you is conscious? I can randomly turn off various neurons of you and check. What atom of a piece of metal has the temperature of 20C?
          1. tatjam · · focus · HN ↗
            This is a good take, but where does one consciousness end, and another begin?

            Does my consciousness extend only to the central nervous system (and peripheral nerves and such are inputs)? Or am I my brain + all nerves? What about organs, does my consciousness include my liver? Is my liver another conscious entity? Going the other way, if I'm using a computer, does my consciousness extend to the computer? I use a computer for well over 8 hours a day, I can use vim motions more naturally than, say, move my earlobes or control my toes. Does that make it part of my consciousness?

            Sadly, I don't think any answer to these questions can be achieved in our current times, or maybe ever if consciousness is somehow non-physical. So even the very reasonable claim that consciousness is an statistically defined property (like temperature) is too far fetched for our current knowledge.

            1. hardbass · · focus · HN ↗
              Indeed thats the difficulty. I used to think potentially conscious AI systems are so far into the far future scifi that I never have to bother, but unfortunately real life forced me to think of these much sooner than I wanted. Its no longer quite a purely academic question now anymore due to the potentiality of current AI being conscious.

              One thing however we can be sure of, "supernatural" stuff is an extremely low to near impossible probability given current knowledge and the fact that history has always won the bet against supernaturalism throughout the centuries.

          2. AlotOfReading · · focus · HN ↗
            There's also the wrinkle that consciousness might be a property that isn't reducible to some defined subset of the system. Imagine we have a haystack in a field. I can remove one piece of hay, and what we're left with remains a haystack. But repeat that process enough times and we'll eventually have something that's not a haystack, even though there's no unambiguous point where removing one bit of hay transforms a haystack into not-a-haystack.
            1. hardbass · · focus · HN ↗
              That's not a problem. Physics has dealt with macro and statistical properties all the time, indeed that was the "wrinkle" that upended physics completely in the late 19th century and led to our modern theories. The thing you said can be said about temperature too. Which atom contains "20C" in a piece of metal? Its a statistical property of a large number of atoms or molecules, a concept that is the bread and butter of modern physics.
Open on Hacker News to reply ↗

Unofficial Hacker News client; not affiliated with Y Combinator.