4 ms·

> The better they get at predicting what the low animals on the surface are doing, the more opportunities that they have to sneak in and sneak out safely to get

by Swizec 2y ago

> The better they get at predicting what the low animals on the surface are doing, the more opportunities that they have to sneak in and sneak out safely to get a meal The most compelling explanation for bird intelligence I’ve read[1] argues that it all stems from social needs. Birds, you see, form lifelong pairs. But they constantly cheat on each other. Keeping track of this cheating behavior, deceiving each other, hide their actions, predict what other birds know, understanding who will and who won’t rat them out to their partner, that’s why the intelligence developed. Then once you have intelligence, it proves useful for all sorts of things. Many species of bird also use this advanced ability to keep track of who knows what for food. They’ll hide a stash for winter and find it all later. But it’s easier to remember where your friend hid theirs than to get your own. So a whole arms race of deception developed. [1] The Genius of Birds https://www.jenniferackermanauthor.com/genius-ofbirds https://www.jenniferackermanauthor.com/genius-ofbirds

reverendsteveii 2y ago
Funny, I've read the same thing about intelligence, particularly trainability, in both rats and dogs. It's driven by a social aspect. The existence of a peer group necessitates imagination and theory of mind: you have to be able to think about what someone else is thinking about. From there you can have thoughts like "That big monkey wants me to do this thing that we do sometimes and has indicated that they'll give me a thing I like if I do it."
- ferguess_k 2y ago
  I wonder if there is any kind of bias of perception. If we agree with this point, this probably means bankers are much more intelligent than engineers, considering the former profession has to communicate with hundreds, maybe thousands of clients throughout their career life, while engineers mostly communicate with machines -- they do communicate with their colleagues but you can see the vast gap between the two. Maybe there are two types of intelligence -- versus humans and versus nature.
  - wat10000 2y ago
    It's a lot more than just two.
  - marcellus23 2y ago
    But a banker and an engineer are 2 individual members of the same species. Evolutionary pressure to develop a specific type of intelligence doesn't apply. But yes, there are of course many different kinds of intelligence.
  - desiderantes 2y ago
    Please go back to your resting place, Mr Lamarck.
  - smallmancontrov 2y ago
    The obnoxious pro-engineering equivalent of your statement is to say that engineers are more intelligent because they are held accountable to a higher standard of objective truth, one that is judged through the cold eyes of physics and math rather than the rosy glasses of a strategically cultivated in-group. It's probably wrong to directly compare these two types of specialization. However, we do have an interesting social experiment going on: China's bureaucracy leans towards engineering backgrounds while the USA bureaucracy leans towards legal backgrounds. You can see this in the strategies pursued by each side: China pulls large and small levers to acquire hard power (in the sense of manufacturing capacity, not just guns) while the USA has historically been better at pulling large and small levers to acquire soft power (and even though tension from the Triffin Dilemma is peaking again, that's still probably a fair assessment). The next decade will probably see a showdown that supports or repudiates the "engineer primacy" vs "lawyer primacy" narratives on the level of international strategy, even though both will obviously still exist and have primacy within their respective niches. Interesting times.
    - sudoshred 2y ago
      Soft power is essentially a euphemism for influence, which in the best case is necessary but not sufficient.
    - 2y ago
- kkylin 2y ago
  Maybe. But at least in mammals, visual information processing is a very expensive operation that involves coordinated activity across many cortical areas. Vision being such a central sensory modality for birds, I do wonder if it was a strong driver for the evolution of their brains. Edit: don't mean to imply a contradiction with the social interaction hypothesis -- needless to say there can be multiple factors that drive evolution in the same direction...
  - reverendsteveii 2y ago
    I just want to agree with you wholeheartedly that there are certainly multiple independent factors driving intelligence. After all, one of the most asocial groups one can imagine, the octopodes, are also at the forefront of animal intelligence (and to further your theory, more of their brains are dedicated to visual processing than we see in humans)
red75prime 2y ago
The sky is the limit in the intelligence race when adversaries are of your own species. Or metabolic constrains are the limit.
scotty79 2y ago
I firmly believe that it's the root of all run-away intelligence. Social species trying to outsmart itself. Predator-prey or individual-environment dynamics are just too slow to compete with self-referential feedback loop. Given that we probably won't need to worry about AI getting significantly smarter unless we put it with existential competition with itself.
- qzw 2y ago
  AI companies are definitely pitting their AI models against both in-house and external models. Didn’t Alpha Zero get to some insanely high level in go by only playing against itself without any reference to existing go literature? So if this generation of AI plateaus at some point, lack of competition again other AI won’t be the reason.
  - scotty79 2y ago
    I don't think it's just about comparing or letting AIs compete. It's more about making success of one AI dependant on it understanding how other AI works (that's uncooperative and has its own goals).
    - sdenton4 2y ago
      Hold on, I've got a great pitch for an AI startup...
yieldcrv 2y ago
Given that every evolutionary outcome is about which genes survived long enough to produce viable offspring that have some of those same genes A) at all B) more than others C) happened to pass on this makes more sense, since its about sex every other use of the same skill is happenstance for example with C), in humans, cancer occurs mostly after sexual reproductive periods, so there is no way for that to have been weeded out of the population. Demonstrating that many non beneficial traits pass on alongside beneficial ones and its all happenstance.
- archimedes237 2y ago
  It can if having non-dead grandparents gives an advantage before reproduction occurs. Could be why we live longer than most species too.
munificent 2y ago
Some species have properties that seem way beyond what would be necessary for evolutionary success. For example, an inland taipan has enough venom in a single bite to kill 100 people. There is no imaginable situation where a taipan is going to need to bite an entire village and take them all out. Why on Earth did it evolve such insanely strong venom? When you dig into it, these outliers are often the result of an evolutionary arms race [1]. In teh case of the inland taipan, they are often prey for mulga snake and perenties, who have evolved immunity to their venom. So you've got a feedback loop where the taipan keeps evolving stronger venom to fight back against predators, who continue evolving stronger immunity to the same venom. Run that loop a million years and you get a snake who can kill a busload of people. Human intelligence is another such outlier. I know it's popular to talk about how animal intelligence is underestimated but even so, human intelligence is just astronimcally greater than any other species. Sure a squirrel can find a bunch of nuts it buried. Humans have built machines and landed them on other planets. Our intelligence is orders of magnitude greater than any other species. My pet hypothesis for years is that this must be the result of an evolutionary arms race within the species. Humans are a profoundly social species. We are mostly too fragile to survive in the wild on our own. The functioning survival unit of humans is a hunter-gatherer group. We are sort of like eusocial animals like ants. But unlike ants who can mostly rely on simple chemical signals to tell which other ants are part of their anthill, we have to rely on social cues. There is a very strong incentive to be able to deceive humans in other groups and infiltrate or sabotage their group. If you're smart enough to sneak in, you can steal a lot or do a lot of damage. Likewise, there's an equally strong incentive to be able to suss those bad actors out to prevent them from doing that. The smarter you are, and the better you're able to remember people and describe them to others, the harder it is to get taken advantage of. Turn that evolutionary crank a few hundred thousand years, and you get a species so smart that the only other animals that can possibly hope to compete with them in terms of intelligence are other Homo sapiens. If we weren't so deeply social, I don't think we'd be so smart. We have these huge brains in order to navigate the fantastically complex social world which we have in turn created by having these huge brains. [1]: https://en.wikipedia.org/wiki/Evolutionary_arms_race https://en.wikipedia.org/wiki/Evolutionary_arms_race
- hanjeanwat 2y ago
njarboe 2y ago
Darwin saw sexual selection as so important to evolution and that he discusses natural selection and sexual selection as different types. Of course sex is part of nature.
RealityVoid 2y ago
They... What? They tell on each other about cheating? Is that actually real? I am very surprised, it seems the kind of communication that is pretty complex, not the kind I expect birds to be capable of.
- Swizec 2y ago
  They do according to the source I linked. At least some species seem to, it's not all birds. You have to remember that "bird" is not like "human", it's more like "mammal". Lots of variation between species :)
yencabulator 2y ago
> Many species of bird also use this advanced ability to keep track of who knows what for food. They’ll hide a stash for winter and find it all later. But it’s easier to remember where your friend hid theirs than to get your own. So a whole arms race of deception developed. This reminds me of a backyard squirrel anecdote from a decade ago. A squirrel would laboriously dig a hole and hide a nut in it, cover it with soil, and tamp down the soil -- until it noticed another squirrel watching it from the top of a fence, at which point it immediately proceeded to dig up the nut and carry it away.
elihu 2y ago
"Lullaby, lullaby, swindles and schemes. Flying's not near as much fun as it seems." -- Peter S. Beagle, The Last Unicorn
gcau 2y ago
How does this theory show that birds didn't first become intelligent, and then started cheating because they're now intelligent enough to do it? The "sitting and watching" theory makes a lot more sense to me.
- ASalazarMX 2y ago
  Also, "cheating" is common in many species, including other birds, and it doesn't seem to affect their reproduction enough to justify evolving cheater powers; as spicy as that hypothesis may be. I also agree that it makes more sense that pressure to become smarter is linked to prediction and preservation, as they're fragile creatures. A wounded mammal could crawl inside its nest and heal, a wounded bird is likely a dead bird.
- Swizec 2y ago
  > How does this theory show that birds didn't first become intelligent, and then started cheating because they're now intelligent enough to do it? I recommend reading the linked book from a bird expert who has been studying this for her entire career. I promise it is much better and deeper than my comment here. It links to 266 references (papers) at the end. I just checked

by Swizec 2y ago

reverendsteveii 2y ago

Funny, I've read the same thing about intelligence, particularly trainability, in both rats and dogs. It's driven by a social aspect. The existence of a peer group necessitates imagination and theory of mind: you have to be able to think about what someone else is thinking about. From there you can have thoughts like "That big monkey wants me to do this thing that we do sometimes and has indicated that they'll give me a thing I like if I do it."

ferguess_k 2y ago
I wonder if there is any kind of bias of perception. If we agree with this point, this probably means bankers are much more intelligent than engineers, considering the former profession has to communicate with hundreds, maybe thousands of clients throughout their career life, while engineers mostly communicate with machines -- they do communicate with their colleagues but you can see the vast gap between the two. Maybe there are two types of intelligence -- versus humans and versus nature.
- wat10000 2y ago
  It's a lot more than just two.
- marcellus23 2y ago
  But a banker and an engineer are 2 individual members of the same species. Evolutionary pressure to develop a specific type of intelligence doesn't apply. But yes, there are of course many different kinds of intelligence.
- desiderantes 2y ago
  Please go back to your resting place, Mr Lamarck.
- smallmancontrov 2y ago
  The obnoxious pro-engineering equivalent of your statement is to say that engineers are more intelligent because they are held accountable to a higher standard of objective truth, one that is judged through the cold eyes of physics and math rather than the rosy glasses of a strategically cultivated in-group. It's probably wrong to directly compare these two types of specialization. However, we do have an interesting social experiment going on: China's bureaucracy leans towards engineering backgrounds while the USA bureaucracy leans towards legal backgrounds. You can see this in the strategies pursued by each side: China pulls large and small levers to acquire hard power (in the sense of manufacturing capacity, not just guns) while the USA has historically been better at pulling large and small levers to acquire soft power (and even though tension from the Triffin Dilemma is peaking again, that's still probably a fair assessment). The next decade will probably see a showdown that supports or repudiates the "engineer primacy" vs "lawyer primacy" narratives on the level of international strategy, even though both will obviously still exist and have primacy within their respective niches. Interesting times.
  - sudoshred 2y ago
    Soft power is essentially a euphemism for influence, which in the best case is necessary but not sufficient.
  - 2y ago
kkylin 2y ago
Maybe. But at least in mammals, visual information processing is a very expensive operation that involves coordinated activity across many cortical areas. Vision being such a central sensory modality for birds, I do wonder if it was a strong driver for the evolution of their brains. Edit: don't mean to imply a contradiction with the social interaction hypothesis -- needless to say there can be multiple factors that drive evolution in the same direction...
- reverendsteveii 2y ago
  I just want to agree with you wholeheartedly that there are certainly multiple independent factors driving intelligence. After all, one of the most asocial groups one can imagine, the octopodes, are also at the forefront of animal intelligence (and to further your theory, more of their brains are dedicated to visual processing than we see in humans)

red75prime 2y ago

The sky is the limit in the intelligence race when adversaries are of your own species. Or metabolic constrains are the limit.

scotty79 2y ago

I firmly believe that it's the root of all run-away intelligence. Social species trying to outsmart itself. Predator-prey or individual-environment dynamics are just too slow to compete with self-referential feedback loop. Given that we probably won't need to worry about AI getting significantly smarter unless we put it with existential competition with itself.

qzw 2y ago
AI companies are definitely pitting their AI models against both in-house and external models. Didn’t Alpha Zero get to some insanely high level in go by only playing against itself without any reference to existing go literature? So if this generation of AI plateaus at some point, lack of competition again other AI won’t be the reason.
- scotty79 2y ago
  I don't think it's just about comparing or letting AIs compete. It's more about making success of one AI dependant on it understanding how other AI works (that's uncooperative and has its own goals).
  - sdenton4 2y ago
    Hold on, I've got a great pitch for an AI startup...

yieldcrv 2y ago

Given that every evolutionary outcome is about which genes survived long enough to produce viable offspring that have some of those same genes A) at all B) more than others C) happened to pass on this makes more sense, since its about sex every other use of the same skill is happenstance for example with C), in humans, cancer occurs mostly after sexual reproductive periods, so there is no way for that to have been weeded out of the population. Demonstrating that many non beneficial traits pass on alongside beneficial ones and its all happenstance.

archimedes237 2y ago
It can if having non-dead grandparents gives an advantage before reproduction occurs. Could be why we live longer than most species too.

munificent 2y ago

Some species have properties that seem way beyond what would be necessary for evolutionary success. For example, an inland taipan has enough venom in a single bite to kill 100 people. There is no imaginable situation where a taipan is going to need to bite an entire village and take them all out. Why on Earth did it evolve such insanely strong venom? When you dig into it, these outliers are often the result of an evolutionary arms race [1]. In teh case of the inland taipan, they are often prey for mulga snake and perenties, who have evolved immunity to their venom. So you've got a feedback loop where the taipan keeps evolving stronger venom to fight back against predators, who continue evolving stronger immunity to the same venom. Run that loop a million years and you get a snake who can kill a busload of people. Human intelligence is another such outlier. I know it's popular to talk about how animal intelligence is underestimated but even so, human intelligence is just astronimcally greater than any other species. Sure a squirrel can find a bunch of nuts it buried. Humans have built machines and landed them on other planets. Our intelligence is orders of magnitude greater than any other species. My pet hypothesis for years is that this must be the result of an evolutionary arms race within the species. Humans are a profoundly social species. We are mostly too fragile to survive in the wild on our own. The functioning survival unit of humans is a hunter-gatherer group. We are sort of like eusocial animals like ants. But unlike ants who can mostly rely on simple chemical signals to tell which other ants are part of their anthill, we have to rely on social cues. There is a very strong incentive to be able to deceive humans in other groups and infiltrate or sabotage their group. If you're smart enough to sneak in, you can steal a lot or do a lot of damage. Likewise, there's an equally strong incentive to be able to suss those bad actors out to prevent them from doing that. The smarter you are, and the better you're able to remember people and describe them to others, the harder it is to get taken advantage of. Turn that evolutionary crank a few hundred thousand years, and you get a species so smart that the only other animals that can possibly hope to compete with them in terms of intelligence are other Homo sapiens. If we weren't so deeply social, I don't think we'd be so smart. We have these huge brains in order to navigate the fantastically complex social world which we have in turn created by having these huge brains. [1]: https://en.wikipedia.org/wiki/Evolutionary_arms_race https://en.wikipedia.org/wiki/Evolutionary_arms_race

hanjeanwat 2y ago

njarboe 2y ago

Darwin saw sexual selection as so important to evolution and that he discusses natural selection and sexual selection as different types. Of course sex is part of nature.

RealityVoid 2y ago

They... What? They tell on each other about cheating? Is that actually real? I am very surprised, it seems the kind of communication that is pretty complex, not the kind I expect birds to be capable of.

Swizec 2y ago
They do according to the source I linked. At least some species seem to, it's not all birds. You have to remember that "bird" is not like "human", it's more like "mammal". Lots of variation between species :)

yencabulator 2y ago

> Many species of bird also use this advanced ability to keep track of who knows what for food. They’ll hide a stash for winter and find it all later. But it’s easier to remember where your friend hid theirs than to get your own. So a whole arms race of deception developed. This reminds me of a backyard squirrel anecdote from a decade ago. A squirrel would laboriously dig a hole and hide a nut in it, cover it with soil, and tamp down the soil -- until it noticed another squirrel watching it from the top of a fence, at which point it immediately proceeded to dig up the nut and carry it away.

elihu 2y ago

"Lullaby, lullaby, swindles and schemes. Flying's not near as much fun as it seems." -- Peter S. Beagle, The Last Unicorn

gcau 2y ago

How does this theory show that birds didn't first become intelligent, and then started cheating because they're now intelligent enough to do it? The "sitting and watching" theory makes a lot more sense to me.

ASalazarMX 2y ago
Also, "cheating" is common in many species, including other birds, and it doesn't seem to affect their reproduction enough to justify evolving cheater powers; as spicy as that hypothesis may be. I also agree that it makes more sense that pressure to become smarter is linked to prediction and preservation, as they're fragile creatures. A wounded mammal could crawl inside its nest and heal, a wounded bird is likely a dead bird.
Swizec 2y ago
> How does this theory show that birds didn't first become intelligent, and then started cheating because they're now intelligent enough to do it? I recommend reading the linked book from a bird expert who has been studying this for her entire career. I promise it is much better and deeper than my comment here. It links to 266 references (papers) at the end. I just checked