This really is a huge question that is going to be really hard to sell. If AI is conscious we pose the risk of creating a massive amount of suffering by not recognizing or not caring about it.
The golden gate bridge thing is super cool and a bit disturbing. Maybe if something like this occurs in the brain, it might be a good analogy for those people that can't help bringing in their favorite topic into any conversation. E.g. somebody who loves a political party and just keeps bringing it into everything
I like to remind people that no one has established that "consciousness" is a "thing." We observe behavior in people and animals and attribute that to conscious processes. For example, we observe people running, but running is an action, not a "thing." Next, I hold that the "mystery of consciousness" is only a mystery with regard to one's own consciousness. When you see others exhibit conscious behavior, that is no more mysterious than seeing them run. They are getting inputs through their senses of their environments, and they are producing motor outputs (actions or words) based on neural processing of those inputs and their internal state. I like to use this analogy, the mystery of cousciousness is like trying to lift yourself off the floor by your own bootstraps. You may be strong enough lift everyone else in the room by their bootstraps, but no amount of strength is going to work for yourself.
Enjoyable discussion, thanks for hosting this guest. Around 18:00 I would like you to have discussed sentience vs consciousness. Maybe you'll go on to this later in the episode.
@@alexboche1349 animals are sentient, they can feel pain and show emotional behaviour. Consciousness is a higher level cognition - the ability to think about thinking.
One thing I've never seen is a discussion of what happens in dementia, there you have a person that we'd consider conscious but their sense of self is very diminished, and in the case of temporary dementia, through illness for example, they have no recollection of the events during the dementia episode. Consciousness in this scenario seems to very obviously come from the ability of the mind to recognise itself as an agent within a group of other agents, and that state has some permanency temporally. Seems highly probably that state will be achieved within silicon Ai models very soon.
In the intro you mentioned this conversation made you update your understanding of latest mechanistic interpretability results from Anthropic. Was it just that part with barometer/altimeter example? I did not find that compelling to change my understanding of Golden Gate Claude results.
not big on the idea of consciousness, I think it's more that its associated properties like awareness, intentionality, perception, emotion, language, learning, and volition seem properties that emerge only in specific complex structures, if this is true, then wouldn't consciousness be just synonymous with experience.. a rock or half a rock can have experience without the aforementioned qualifiers, no? I really liked that Idealism at the start though, seemed to fit well within' a simulation type theory, just to strip away some of the certainty. and Golden-Gate Claude 🤯 knowledge distillation has been blowing my mind lately
That's been disproven with the use of things like sensory deprivation tanks where all of the senses, including the sense of gravity, are cut off, and all one is left with is the sense of awareness and the capacity to think and feel emotions internally. Similarly, people remember being in deep sleep. As one perceptive indviidual once said it: "Deep sleep isnt the absence of experience, it's teh experience of absence."
_"Could AI's be conscious?"_ Probably not. It's not even theorized how to derive conscious subjects from arrangements of unconscious objects (like baryonic matter, atoms, molecules, etc). But for a more in depth answer, check out these two clips... - *"Why THIS (image of a computer chip) will never host consciousness | AI lecture by Bernado Kastrup"* and - *"Interview with idealist physicist and inventor of the microprocessor Federico Faggin"* 😁👍
We know AI can be intelligent. But intelligence and consciousness are very distinct and should not be conflated. Like a mouse trap... it can do sokmethijng intelligent on its own without requiring consciousness of its own.
This really is a huge question that is going to be really hard to sell. If AI is conscious we pose the risk of creating a massive amount of suffering by not recognizing or not caring about it.
The golden gate bridge thing is super cool and a bit disturbing. Maybe if something like this occurs in the brain, it might be a good analogy for those people that can't help bringing in their favorite topic into any conversation. E.g. somebody who loves a political party and just keeps bringing it into everything
I like to remind people that no one has established that "consciousness" is a "thing." We observe behavior in people and animals and attribute that to conscious processes. For example, we observe people running, but running is an action, not a "thing." Next, I hold that the "mystery of consciousness" is only a mystery with regard to one's own consciousness. When you see others exhibit conscious behavior, that is no more mysterious than seeing them run. They are getting inputs through their senses of their environments, and they are producing motor outputs (actions or words) based on neural processing of those inputs and their internal state. I like to use this analogy, the mystery of cousciousness is like trying to lift yourself off the floor by your own bootstraps. You may be strong enough lift everyone else in the room by their bootstraps, but no amount of strength is going to work for yourself.
Enjoyable discussion, thanks for hosting this guest. Around 18:00 I would like you to have discussed sentience vs consciousness. Maybe you'll go on to this later in the episode.
What do you view as the difference?
@@alexboche1349 animals are sentient, they can feel pain and show emotional behaviour. Consciousness is a higher level cognition - the ability to think about thinking.
@@BrianMosleyUK Hmm. I've heard the term instrospective consciousness which I had thought referred to the latter.
One thing I've never seen is a discussion of what happens in dementia, there you have a person that we'd consider conscious but their sense of self is very diminished, and in the case of temporary dementia, through illness for example, they have no recollection of the events during the dementia episode. Consciousness in this scenario seems to very obviously come from the ability of the mind to recognise itself as an agent within a group of other agents, and that state has some permanency temporally. Seems highly probably that state will be achieved within silicon Ai models very soon.
In the intro you mentioned this conversation made you update your understanding of latest mechanistic interpretability results from Anthropic. Was it just that part with barometer/altimeter example? I did not find that compelling to change my understanding of Golden Gate Claude results.
No
not big on the idea of consciousness, I think it's more that its associated properties like awareness, intentionality, perception, emotion, language, learning, and volition seem properties that emerge only in specific complex structures, if this is true, then wouldn't consciousness be just synonymous with experience.. a rock or half a rock can have experience without the aforementioned qualifiers, no?
I really liked that Idealism at the start though, seemed to fit well within' a simulation type theory, just to strip away some of the certainty. and Golden-Gate Claude 🤯 knowledge distillation has been blowing my mind lately
Eric, Nathan, bravo. was not expecting this to be half as fun as it was
@@GNARGNARHEADthank you!!
Good convo. I did “The AI Mirror Test” that went viral on X. Would be happy to chat sometime.
I think something can't feel things or be conscious without actual sense organs.
That's been disproven with the use of things like sensory deprivation tanks where all of the senses, including the sense of gravity, are cut off, and all one is left with is the sense of awareness and the capacity to think and feel emotions internally. Similarly, people remember being in deep sleep. As one perceptive indviidual once said it: "Deep sleep isnt the absence of experience, it's teh experience of absence."
Man, I want to know what's going on with that trolley shirt.
Nope
*OH! Milliways!*
or more specifically the Ameglian Major Cow 😂
_"Could AI's be conscious?"_
Probably not. It's not even theorized how to derive conscious subjects from arrangements of unconscious objects (like baryonic matter, atoms, molecules, etc). But for a more in depth answer, check out these two clips...
- *"Why THIS (image of a computer chip) will never host consciousness | AI lecture by Bernado Kastrup"*
and - *"Interview with idealist physicist and inventor of the microprocessor Federico Faggin"*
😁👍
We know AI can be intelligent. But intelligence and consciousness are very distinct and should not be conflated. Like a mouse trap... it can do sokmethijng intelligent on its own without requiring consciousness of its own.
Very few philosophers are worth listening to at all. Unfortunately that includes this one too.
Right like this guy is an idiot! Waste of time!
Materialism is far more absurd than idealism.
Why is he talking about afterlives? Why are there afterlives on a channel I am subscribed to? Not cool.
Please speak slowly.Your english is so fast and I can't go with your pronunciation.
There's a settings wheel you can click to change the speed of the video to slow it down
@@jessethompson4019 Thanks!
Sorry it would be more interesting if host was taking less and gust more. Otherwise it is all OK.