Not to discount the research or its usefulness, but I feel uncomfortable with sterile language that, to me, implies that humans are nothing more than accidental automatons that can eventually be replicated and replaced by AI.
> Not to discount the research or its usefulness, but I feel uncomfortable with sterile language that, to me, implies that humans are nothing more than accidental automatons that can eventually be replicated and replaced by AI.
>
> Aren't we more than that?
Are we more than that? What evidence do you have, aside from "God loves us"?
You are carrying around an amazing machine in your noggin, but I haven't seen any convincing evidence that it can't be duplicated. As far into it as we have seen, it is all laws of physics, chemistry on top of that, and biology on top of that.
We definitely haven't cracked the "self learning algorithm" yet. You can show a person a couple pictures of a dog, and they can construct a 3-D mental model of a dog, and what it looks like from all angles and in all poses. Yet you have to give thousands of examples to do the same thing for a ML/DL system.
How much of the ability of those models is baked into the genetic code of humans? How much is learned as a toddler playing with blocks to figure out basic physical properties like object permanence, gravity, friction, and such? By just watching and interacting?
> You can show a person a couple pictures of a dog, and they can construct a 3-D mental model of a dog, and what it looks like from all angles and in all poses. Yet you have to give thousands of examples to do the same thing for a ML/DL system.
By the time a human brain can do this, aren't they also exposed to thousands of examples? How do we learn to speak for example? It's a long painful process of trial and error. How is this significantly different from a supervised learning algorithm? Teaching a child to catch is very similar. At first, they can't recognize the trajectories of thrown objects very well and they aren't coordinated enough to move their hands to where the projectile is. After time and many repetitions, they get to the point where they can catch a ball without consciously following the ball or can even catch without looking based on initial trajectory.
What makes you so sure an "artifical" system has no qualia?
Do you find that hard to believe? Why?
Why could having qualia not be some kind of inherent, perhaps emergent property of all systems complex enough to be called intelligent? And why is this an issue at all? This could be some unknowable mysterious meta-physical property that has no practical bearing on anything.
What is the angle here? Are you worried it is important? There is literally zero indication it is important or am I missing something here?
Personally, the more I learn about the intelligence of myself and others, the less I believe that we are more than a bunch of pattern recognition layers.
It used to haunt me. I have come to acceptance. Of course I could be wrong, but that’s my intuition at this point.
Do you experience subjective existence or are you a non-conscious entity? If you are a set of pattern recognition layers responding to stimuli, then I would not be so shocked you are unconcerned about the hard problem of consciousness.
I have basically fallen to the side of the Copernican principle on this.
Consciousness feels like it may at least partially be our fancy word for “soul” in the 21st century.
Measures of basic intelligence appear to us in crazy places in the animal kingdom. We will find it is not special at all.
I believe we are nothing but of a bunch of chemical states, the patterns and flow of which could be replicated in another system, natural or artificial.
On this planet, we just have the fastest, most complex and efficient processor going at this time.
Edit: I should add there may be more to human intelligence than “just” patter recognition, but that seems to do the bulk of the work.
> On this planet, we just have the fastest, most complex and efficient processor going at this time.
Not neccesarily. We haven't really figured out how to benchmark other species' brains properly and I wouldn't be surprised if something like individual bottle-nosed dolphins or elephants couldn't rival at least some individual humans.
> If you are a set of pattern recognition layers responding to stimuli, then I would not be so shocked you are unconcerned about the hard problem of consciousness.
I am as unconcerned about the hard problem as I am about elan vital, because there is just as much "compelling" evidence for either.
Perceptions taken at face value are not evidence. Evidence is an observation interpreted within a consistent body of knowledge. Our consistent body of knowledge is inconsistent with the naive perceptions of qualia.
You can take the position that qualia are direct evidence and so our body of knowledge is mistaken, or you can take the position that qualia are likely deceptive and so cannot be taken as direct evidence, just like all of our other perceptions.
I'm constantly amazed how many people think the former position is more plausible.
I walk around on legs. I can’t be sure if I am moving because of the legs or through some other action. Some people who dislike the idea of legs are telling me I don’t really have legs.
If you also claim that your legs let you walk on water and sometimes even walk off a cliff without falling, then I'd agree your analogy is in the right ballpark.
So then we agree, you are saying your legs exist and are magical and therefore our understanding of the world must be wrong, and I am saying that our understanding of the world implies that magic doesn't exist and so you must be mistaken about the magic of your legs.
No, I am saying the analogy is that my legs exist and may or may not be magical and therefore our understanding of the world must be incomplete because it has no explanation for legs in the first place magical or otherwise, and you are saying that our understanding of the world implies that magic doesn’t exist and so I must be mistaken about the magic of my legs and probably their existence as well.
Saying "our understanding of the world is incomplete because it has no explanation for qualia" is a classic, fallacious god of the gaps argument.
In fact we do have the beginnings of viable neuroscientific theories of consciousness, including qualia [1], thus showing that the claims of a hard problem were just the same old smoke and mirrors we've seen time and time again in claims to human specialness, just like with vitalism.
> and you are saying that our understanding of the world implies that magic doesn’t exist and so I must be mistaken about the magic of my legs and probably their existence as well.
Nah, our only disagreement is over the magic. I don't deny the perceptions of qualia, simply their special, "ineffable", "experiential" character that you implicitly conclude they must have. That's the magic that doesn't fit into our body of knowledge and that's the only "hard problem" of consciousness, a problem that doesn't really exist because it's based on erroneous assumptions, as described in the paper above.
Just like placing a pencil in a glass of water gives the illusion that the water has broken the pencil, so your introspective perceptions trick you into erroneous conclusions about your awareness.
But what exactly is the erroneous conclusion here? It clearly exists widely and many people report having it. We don’t know how it works or its mechanism. We have no idea how to create more of it outside of having babies. We have no idea if animals have it. A recently dismissed Google employee was unable to prove to himself that a chatbot didn’t have it. A physical explanation would require completely new science as science has almost no existing concept of the phenomenon.
> But what exactly is the erroneous conclusion here? [...] A physical explanation would require completely new science as science has almost no existing concept of the phenomenon.
That is the erroneous conclusion. "Qualia", whatever they are, fits squarely into physicalism and there exists literally no evidence that anything more is required. The paper I linked is proof that qualia can be accounted for within science as it exists. Again, you're asserting a god of the gaps.
> It clearly exists widely and many people report having it.
That doesn't mean that water actually breaks pencils. The meaning of any observation is always interpreted within a consistent body of knowledge. The problem is that people really want to take the perception of their experience as a direct observation of reality that doesn't need to go through that interpretation step, and that's simply incorrect.
What AI doesn't have is access to this world. Instead it has a fixed training set. The world is a much richer, dynamic training set. So AI would need an android body and to become part of human society to experience the real world as we do.
In my view human intelligence is being cracked open like a nut as we speak. It's a matter of decades - honestly, I think years - before all kinds of shit will hit all kinds of fans. However, irrespective of the societal impact, intelligence is very interesting and there is a lot to be gained from decoupling it from the human form.
I've handled this by looking at the (mis)usage of language. "Automatons", "pattern recognition", "statistics". These are words. And they all carry with them a palpable sense of dismissal, of a certain arrogance that I think is misplaced. Note: If you use these words fully egoless and objectively, great, that is excellent, but I don't think most do.
If you'd start to regard intelligence in it's abstract, non-material, form as a "divine" gift to all "living" beings the picture starts to change. There is a very real, very deep sense of wonder to what is happening that gets lost whenever you call intelligence - biological or digital - "just pattern recognition" or something to that effect. You reduce the wonder that is intelligence to something you hold in your head as simple and reductionist. It is in that very moment you diminish it and make it non-interesting. It's like calling human embryos just a bunch of cells. It's not wrong per se, but it's said in a way that reduces the complexity of the system to ridiculous levels. Even without invoking metaphysical/religious beliefs you can imagine an embryo being a highly-complex physical system with astronomical levels of emergent properties that are effectively incalculable, for us at the moment if not at a fundamental mathematical level.
If you look at machine intelligence like perhaps a small child would, full of awe and wonder, I bet your outlook on things can change. Humans are carriers of this "intelligence", they are not optimal and they are not alone. Who cares? Do we need to be the sole proprietors of intelligence in all its forms and shapes in order to be happy? I say enjoy the ride.
I don't think so. I think that we're a brew made of very simple ingredients that results in something that strives to live and reproduce. That's the main missing ingredient in AI for me; desire, or preferences. We just show them a bunch of things in turn, and tell them if they're right or wrong with some reduction of those things. They don't even desire our approval, we've built systems that process, they don't desire.
I think it's because we've nearly figured out cognition (which is ultimately simple), but not the desire to live or to reproduce (which as people we half the time think of as sinful desires), which we probably have deeply hardcoded within ourselves. The desire to emulate others derives from those more fundamental desires (without which we each wouldn't have existed in an unbroken lineage from the first cell), because we can look at others as experiments to model ourselves after, or when we see them fail, make guesses as to how they may have been mistaken in their choices.
We're physical things. The really mysterious part of our situation imo is how our perspective (for lack of a better word) attaches to our bodies (and minds.) Why am I me and not you? Why am I physically located in this mind? Would a perspective attach to a machine with cognition and desires? Would there be any physical way to prove that one hadn't, so would we morally just have to accept a machine's experience as similar to ours? We can't prove other people aren't machines, so we (generally, with extreme difficulty with people very unlike us) just take them at their word.
That's actually a reason I'm against full AI, because it involves creating a bunch of machines that we wouldn't have the moral authority to shut down. Especially if we haven't refined the technology and it's just as power-inefficient as what we have now. Sparking neural networks where electricity isn't used for the signal, just as a potential within the cell walls, and the spark is a little cough of neurotransmitter that is immediately sucked back in. That's just unbelievably low power (even though iirc it's still half or more of the calories you eat.) And brains are so space-efficient. I could cram 100 people in a room full of one AI, and the 100 people would consume fewer calories.
We're accidental automatons that are very cool and very difficult to reproduce, and still very mysterious in a fundamental way.
I'm "getting" to watch a couple close relatives go through severe cognitive decline.
From what I've seen: no, we're not.
They frequently drop the last few minutes of memories on the floor and you can watch them reset, ask the same questions, react the same way, all in precisely the same tone as the first time. Over, and over, and over. We're little more than spinning tops—they've a condition that makes that plain, but I don't see why the rest of us would be any different. We just retain and re-apply more of our input, than they do, so it's less obvious that "the lights are on, but nobody's home".
There is no behavior in our universe that exists outside the laws of physics. So whatever we are, it is a product of our physical form. If that's true, then we can build something else that also exhibits our same behavior but is a purely non-biological organism.
You've made the assumption that the laws of physics allow for something that also exhibits our same behavior but in a purely non-biological organism. I see no reason why this must be so. The laws of physics are not entirely understood and neither is the brain.
Even granting your premise that the laws of physics allows for such a thing you've also made the assumption that humans are intelligent enough to "built something else that also exhibits our same behavior but is a purely non- biological organism". You presumably don't think a dog could invent such a thing but are assuming evolution has given humanity the ability to do so.
Agreed. In a sense, that is kind of my point: Can physicists explain in extreme detail how the mind physically implements omniscience (instances of which are on display in this thread)?
Your “laws of physics” are just a verbal representation of what has been observable to you. It’s not a fundamental property in the deepest sense and there’s no guarantee that they may not change.
It's not a fundamental property in the deepest sense
What do you mean by "in the deepest sense"? If the actual laws of physics were to change, then our world would instantly change - perhaps in such a way that we no longer can exist. If our understanding of the laws of physics were to change, that doesn't undo what we already know. For example, apples won't start falling up from trees. Greater understanding of the laws of physics refines our knowledge, for example what, exactly, happens inside a black hole? As far as the physics governing our daily lives and even our biology, it appears to be rather mundane and well-known. It's not apparent that resolving QFT with GR is going to answer the question for how life arises.
I think it therefore stands that we could, in principle, build something equivalent to a human being - I just see no reason why we would choose to do so. After all, we're already pretty good at building new human beings!
Can you solve a system of coupled Schröedinger equations and know what my wife will want for dinner? I’m only partially joking of course. What I mean is that physics conveniently formalizes what is most conveniently formalized, thus giving an impression of all-encompassing knowledge and predictive ability, when in fact it’s only a small fraction of what is. And I’m saying this as a PhD in chemistry.
Do you believe consciousness doesn't fall under the laws of physics or are you saying the laws of physics as currently known don't yet go far enough to explain the working of consciousness? You probably know Roger Penrose and Douglas Hofstadter have been going round and round on this issue, in a good-natured manner, for decades now. Penrose is convinced QM is required to explain consciousness, whereas Hofstadter maintains consciousness can be modeled with mathematics. It really boils down to can you create an AI that's conscious (which Hofstadter believes) or do you need a physical structure like the brain (which Penrose believes). Neither one is arguing though that consciousness isn't subject to the laws of physics, one just believes the laws of physics as currently known are insufficient to explain consciousness. That appears to be the camp you're in?
Flip it around. Are you saying that you'd be comfortable with the idea that human behavior is completely and utterly random and unpredictable? I'd like to believe that every person has the ability to steer their own future (within the extent that their circumstances allow) and those of their families, and that requires a certain level of pattern recognition and predictability for that to succeed.
And anything that can be calculated by a human can possibility be calculated by an AI as well.
If someone asks you if you want vanilla or chocolate ice cream, once you have made a choice, even dualists would agree it is just a dance of chemistry that sequences a series of neural firings which propagate, excite muscle contractions, resulting in the verbalization of "chocolate". That is, the soul isn't acting at the level of muscles or plucking the strings of individual neurons to make them fire.
Instead, dualists have a vague assertion that the soul influences (perhaps determines) the choice itself. Getting to my point: if you believe that the soul drives the choice and the decisions wasn't just the result of billions of (purely physical) neural antecedents, where in your brain are the rules of physics violated in order to steer that neural outcome we call "choice"?
One hand wave is to say: quantum! Yes, quantum effects are not individually predictable, so one could imagine that a tricky soul which is trying to not be discovered somehow can cause just the right statistical fluctuation in the right neurons at just the right time to nudge the outcome to what it desires. But consider that that activation energy of a synapse is about 30 orders of magnitude greater than quantum fluctuations at the atomic level. Or put another way, not every butterfly flapping its wings causes a hurricane. Such things happen only if the interconnected decision points are on a knife's edge. It seems like wildly overconstraining to think that our brain configuration must be so delicately balanced for every decision we make to allow a soul to have the opportunity to effect every nudge that is required of it.
Taking a step away from humans, spiders have pinhead "brains". They can navigate the world, choose a place to build a web, build a web, catch prey, avoid harm, repair their web, mate, reproduce. I think few dualists would credit spiders with having a soul. Why is it surprising that a brain a billion times bigger is capable of making music?
Nope - the material universe seems to be able to achieve unimaginably spectacular feats (e.g. us) within the confines of physics, chemistry, and biology. I won't be surprised if such creativity will find a way to port "consciousness" across substrates (from meat to metal).
But I don't see this fact as dismissive of humanity's grandeur - the idea that we were placed here and made conscious by something outside of creation (quite literally a deus ex machina) is clunky. How much more elegant is it that we're instead of a kind with the rest of the universe around us and a link in that spectacle?
It's easier to start tackling the question when you make the distinction between mind and consciousness as other commenters pointed out in other threads. Let's just assume some very basic definitions:
- mind = memories, personality, aspirations, knowledge, pattern recognition, etc...
- consciousness = this unmistakable, unshakable certainty that "I exists, I am" that we all share (unless solipsism is true and you are a philosophical zombie :) )
To me it is then evident that the answer to your question is: we (and any sentient being for that matter) are more than that any AI will ever be.
I wonder what the venn diagram looks like between “people who are responding to this in the negative” vs “people who believe human rights are real and serious.”
I’m not trying to be flippant here, I just really don’t believe that people think through the consequences of their beliefs very well.
In case you’re interested, in the 20th century James in his pragmatism essentially did this for all of philosophy. He divided philosophy in two 2 main types, roughly the pessimist and the optimist or tender method and ____ method (forgot the second), and said all of philosophy is due to these personalities. Dewey made this Venn diagram.
I don't think humans are particularly special, and I don't believe in human rights. At least in the sense that there is some fundamental guarantee to liberty, property, or whatever.
Nature is cruel, and if given half a chance, will for example snatch the life from a 2-month-old baby. That baby did nothing wrong, but did not have the "human right" to life and liberty, just a hope for it.
We can still institute rules we agree to abide by, in hopes that our lives will be better overall. But that's just self-interested agreement, not some immutable "right" inherently bestowed upon us.
> I don't think humans are particularly special, and I don't believe in human rights.
I don't believe humans are particularly special, but "human rights" are nothing more than what societies deem as requirements for leading fulfilling lives and the degree to which we have those rights is determined solely by the willingness of society to protect those rights. There are no "natural rights" only rights that we recognize and implement for ourselves at a societal level. There is nothing stopping a society from making access to something arbitrary a "human right" like the "right to be delivered a new blue hoodie every winter solstice" and then protecting and enforcing that right. If a society decides that humans have no rights, then you'd be correct in that there are no human rights which is what you may experience in some parts of the world.
When you sum all the accumulated knowledge about the world and the universe, it seems to me that us being nothing more than accidental automatons is the most likely scenario.
Aren't we more than that?