You're applying evolutionary forces to a simplistic binary plateau.
This isn't intelligence, it's a subset of maximal solutions to bounded problems.
Hint: your dog bounds you as much as you bound your dog. Your dog has some input into your modern homo sapien sapien mind. Same goes for your atmosphere, your gut fauna, your planet's iron core and so on.
By following this logic, would Dogs evolve to be more and more intelligent, given that we feed them nothing but a high-calorie diet with zero need for any survival skills? I suspect that high intelligence does not give them any evolutionary advantage and therefore wouldn't become dominant, but then humans do like "smart" dogs and could selectively breed them for intelligence. No?
I like to think of AI-Genesis through the lens of what humanity has already done through domestication. We take something primitive and progressively adapt it to serve a greater utility. I think working dogs are the most interesting example of this. We've taken a species, the wolf, and made it smarter while also making it want to do work, learn tricks, and follow orders. Of course, you still need to train the animal for optimal results but even breeds like collies know how to herd instinctively.
Anyways.
Let's assume the best and brightest dog breeders endeavor to make German Shepherds as intelligent as they possibly can. Would the same ethical debates about what constitutes a 'mind' come into play? What would happen if the dogs became smart enough to make their own mating decisions? Would we be worried about them turning on us once they get close to human level intellect? Would it be immoral to make these dogs work? Or, would not letting them work be considered immoral?
This is just food for thought. But I suspect AI's capabilities will grow much in the same way other domesticated species have grown into the specialized roles we've crafted for them.
This brings up for me an interesting question. Does the rise of intelligence in one creature put an evolutionary pressure towards it in other species, and how rapidly can that manifest. We’ve certainly bred dogs for intelligence, border collies are genuinely amazing creatures and that developed in a somewhat human time scale.
There seems to be more dimensions to it though. Trivially, a dog and a human might be on one axis, but you can't rank differently intelligent humans along that same axis. The average dog, while strictly and obviously not as intelligent as a human is not impaired, they function at their fullest. While as humans go further down and away from mainline human intelligence become obviously impaired.
We seem to have a notion that intelligence has multiple dimensions to it, so it's more likely that human intelligence exists as some kind of probabilistic gradient on a hyperplane in a high-dimensional space, and sampling from that plane yields any possible human intelligence profile, while sampling away from that space yields something decidedly not human. Other animals thus must be similar hyperplanes, perhaps in their own space, and likely not intersecting with humans at all.
I'm actually not stretching it, I was demonstrating that your limitations either exclude humans or include animals.
What your limitations seem to ignore is that not only does most of our day-to-day behaviour fall under what you consider instinctive, but also that you seem to be ignoring the fact that many animals communicate in languages you might not necessarily be able to understand (Dogs, for example, communicate mostly in body language, looks, etc.).
Can you state an example of a complex thought? I can think of a fair amount of thoughts I would use to denote intelligence that would fall under what you consider 'instinctual'.
I'd like to see someone make the argument that current models aren't just combining a number of "tricks", similar to a trained animal. My dog can "sit", "stay" and "beg", all using the same model (its brain). Is the dog generally intelligent?
He claims humans can't increase the intelligence of animals, but we probably can?
Dogs probably have higher emotional intelligence (by human standards) than wolves, and we probably could breed dogs (or have we?) that are smarter at basic puzzle solving activities like opening latches with their paws or whatever.
Even dog-like intelligence would be extremely useful.
It's interesting when training a dog. Sometimes you can see the gears turning and know their thought process and conclusion before they do, since we are just much smarter. And other times, they come up with something that makes no sense at all based on their "dog logic". I guess their brain just works differently.
Anyway, focusing on simply humans may be short sighted. I often understand systems better by comparing them to similar systems and working out the differences. If anything bears fruit I think it'll be the researchers that are starting small and trying to replicate a worm brain. Then build from there, faster than evolution, because we can.
Not really - genetic bottlenecks aren't a feature of intelligent animals, it's a feature of a population that's undergone some kind of very dramatic pressure event like catastrophic loss of habitat or mass illness.
Bottlenecks are more likely to be an issue for species with a limited range, population and a slim range of survivable conditions. Intelligence has a way of expanding those ranges by making you more equipped to access otherwise inaccessible resources. Think of elephants navigating between watering holes during droughts, crows using tools to fish out hard to get food, or orcas using sophisticated hunting techniques to get at prey they otherwise could not.
Intelligence adds flexibility, and flexibility is especially useful under pressure.
You're still anthropomorphising too much by comparing it with an animal. A large amount of an animal's behaviour is genetically determined, wrought by Mother Nature step by step in thousands of ad-hoc adaptations, and we simply don't know what many of those adaptations are. They come together in something of a cohesive whole, each part depending intricately on all previous parts, and that's why we socially understand dogs to the extent that we do: lots and lots of the machinery that makes us both is shared.
By contrast, hopefully we'll design our AIs according to, say, engineering principles. That way, we can know what its motivation is, which I'm sure you'll agree is very important when creating something that is potentially substantially smarter than us. This is so unlike any other intelligence we interact with, that there's no particular reason we should be able to compare it to them.
Of course they're not... Dogs are one of the smartest creatures we know of. Dogs and humans might as well be considered the same level of intelligence compared to something like an LLM.
It would be more interesting to hear the discussion of a more realistic comparison. What about a fly? A Venus fly trap? A dandelion? A river flowing downhill?
Those feel more in the absolute best case neighborhood when considering the potential "intelligence" of the current state of AI.
Our understanding of what intelligence even is is so limited we barely have the right words, or "thinking tools" as Daniel Dennet calls them, to even have this discussion.
Compared to animals sure. But we are the first thing to evolve intelligence, it's unlikely we are anywhere near optimal at it. It'd be like the first amphibian believing it was good at running.
True. There's a theory that says that intelligence evolved as a tool to outsmart your pray. So top level predators will always have to be smarter than their pray.
Therefore, survival needs intelligence.
Personally, I think we are going to hit a hard stop on our current path to AGI. We are looking at intelligence from the top level hoping to somehow duplicate it. What we are getting is a dumb mimic with zero intelligence. We need to start at the bottom and build our way up. So, pick a low intelligence animal and figure out how they perceive the world and survive in the world. Trying to mimic intelligence won't get us to AGI.
The thing is, if you take the species dogs, pigs, horses, apes, and humans, then you can clearly put them somewhere on a scale regarding their ability to learn or understand (a) simple language (b) complex language, (c) simple math (d) advanced math. There might be some edge cases, but an average human will be better at any other species no matter how hard you try.
There is a very hard species boundary, so it must be something that is giving to the offsprings.
If you further accept that the ability to learn advanced math does not fall from the sky and we somehow descended from an ancestor that did not have this ability, then the only way we got here is that (something that correlates to) this ability was heavily selected for in the past.
And you need some kind of continuum (it can also be multi-dimensional) where you can gradually converge to. I don't think this is something that spontaneously emerged.
If you accept all this, then the same things will likely still play a role within current humans.
Yann LeCun often argues that animals like cats and dogs are substantially more intelligent than current LLMs [0] and I'd have to agree. I don't see how/why to consider practical knowledge as only constituting a tiny fraction of our overall intelligence. Either way, it's not clear if the GPT-* models will someday produce emergent common sense or if they're going down an entirely wrong path.
The problem with your logic is that improving a humans intelligence by as much as the difference between a dog and a human is a HUGE leap in intelligence. There are far smaller jumps in intelligence that are still meaningful that fit into the span of dog to human intelligence.
Probably not by natural pressures, though. We're more than intelligent enough to outcompete any species that might try to fill the same ecological niche as humanity. No benefit to even more brain power. It would more likely take intentional selective breeding if we wanted to go that route.
Animals certianly aren't ahead of us in human cognitive function, but that feels like a loaded competition. Humans do have a hard time imagining intelligence as anything other than what we have.
My dog has been known to be more perceptive and cunning than humans at times (including me), so I'd have to say yes.
It ultimately depends on the breed, upbringing, the dog's individual personality, and so on. Generally speaking however, there's a mountain of empirical evidence that demonstrates how extraordinary canine intelligence can be.
Humans have a rather long history of discounting other species' intelligence, and it seems as time progresses we learn just how wrong we were. I was reading an article the other day about how turtles of all species are actually much more intelligent than previously thought, in part due to flawed behavioral studies in the 1950s and 60s. They aren't even mammals. It makes one wonder:
I eat meat. I enjoy it. If anything I'm aware of how unethical it is, but I currently rationalize its consumption partly as a tacit acknowledgement of how dark and broken the world is, and partly for nutritional and cognitive health reasons.
Yet, it's still terrifying to think that if you remove the notion of humans being unique in our ability to experience suffering—not only physical pain, but notably psychological pain—we're essentially inflicting what amounts to large-scale genocide of creatures with feelings, on a daily basis. It's suffering on a scale that is incomprehensible.
To circle back to the topic at hand, it's curious that consumption of canine meat is considered highly immoral or more commonly is outright illegal in the majority of the civilized world. Yet, use of other mammals as livestock—mammals who arguably exceed canine intelligence in many instances—is widely accepted.
I think that dissonance underscores just how badly we've failed in recognizing intelligence beyond that of our own species. Intelligence, the capacity for suffering, sentience—it all has been unduly anthropomorphized.
You're applying evolutionary forces to a simplistic binary plateau.
This isn't intelligence, it's a subset of maximal solutions to bounded problems.
Hint: your dog bounds you as much as you bound your dog. Your dog has some input into your modern homo sapien sapien mind. Same goes for your atmosphere, your gut fauna, your planet's iron core and so on.
Your models cannot even touch this at the moment.
reply