Comment by rixed | Hacker News Reader

Comment by rixed | Hacker News Reader

Parent Full thread

I've met a few people with the same amazing optimism that you describe here and it always surprised me (thank you for that by the way, I understand it requires more nerve to display optimism than pessimism as the former is often associated with silliness whereas the later is more often seen as "sad but true").

This kind of reasoning feels like bad extrapolation : you consider humans 100k years ago and humans now, and you extrapolate ; sure enough, it looks like we are in a journey to become some abstract intelligence filling the universe.

But if you take more samples in the so much longer history of life on earth, you will see that among all life forms that appeared and disappeared, human-like intelligence emerged only once, by accident. And it seems to me that the chaotic evolution of human civilization strongly suggest that it will also disappear accidentally.

Natural selection does not select for that kind of intelligence. Awareness of the environment, yes ; good predictor of the near future changes of this environment, yes. But the creation of larger and larger organisms like a civilization that would evade Earth? I see no reason for that in theory and no evidence in practice.

I think we're making orthogonal points.

I agree that life is very unlikely on any given earth-like planet. And most life out there is probably of the unicellular sort.

I also don't think its guaranteed that humanity will last beyond the point we can 3D print an antimatter supersoaker.

I am just saying that out of 50 sextillion chances, I bet enough planets evolved human level intelligence that at least one made it to the singularity. And given that most rolls of the dice happened billions of years ago, whatever the singularity turns into exists already, and has for billions of years.

I don't think this is optimistic in the slightest, I think its a very sober reading of the probabilities. I would actually say that thinking we're unique is the optimistic point of view.

But I think that what I'm saying feels optimistic because its contrarian, and its contrarian because our intuitions just aren't calibrated to reason through this sort of thing. A lot of it is because of our fiction. We imagine green space men because that's what actors dressed up in suits can portray. We imagine aliens like ourselves, don't find that, and conclude there aren't any of any sort. We imagine other civilizations out there to be contemporaneous with ourselves and on the same timeline because that's all we've ever seen. And we can't grapple with numbers like 50 sextillion because we're evolved to count bananas. But if you think about the odds we're the only intelligent life that evolved and just reason through it using numbers, how can you think we're alone? There are only 100ish atomic elements that matter. They're all pretty common. Our star is very average. The Earth is a pretty pedestrian rocky planet. The Milky Way is pretty average. Life emerged relatively quickly after the planet cooled (on an astronomical imescale). The fact that we look out into space and don't see anything means diddly squat because we can only observe 5 earth-like planets/moons out of 50 sextillion. But we feel alone and unique because we're evolved to assume that all we see is all there is.

Ok, thank you for the additional explanations, it does make more sense now.

I do not think we are unique but I believe we "human like intelligences" are a very unlikely product of as many tiny random adaptations as to make the number of rocky planets in the observable universe look like a small number.

This is how I look at it: the Earth is an ecosystem with millions of species that's been evolving together for billions of years, adapting to each others changes as much as to the change in climate etc. The combinatoric is huge, but some environmental pressure make that process invent over and over similar solutions to the same class of issues. Some traits are totally random and I believe will exist only once in the whole history of the universe, and some traits are heavily constrained by environment/chemistry and are invented over and over by evolution, on Earth and elsewhere. Is intelligence of the former or later class of traits?

Some organ to predict the future is certainly part of the heavily constrained invention. But I can't convince myself that a human like intelligence able to escape Earth, both in imagination and in practice, is anything more than a fluke. I just don't see how to logically connect super-intelligence with how evolution works.

On a higher level, your idea also evokes a whole set of similar ideas that are all based on multiplying tiny chances of high intelligence with sextillions of planets, such as "life is a simulation" or "matrix like brain in a jar". My intuition tells me that we can't extrapolate that far without the unknown dominating everything.

This is so interesting to think about. Where do you see the extremely unlikely jump from non-intelligence to human-level intelligence that makes us a fluke? I think paleontology at least indicates that evolving monkey level intelligence is inevitable. We have (somewhat independently) evolved:

crows and parrots dolphins elephants chimps and orangutans octopi

Just looking more into this because is this discussion is fascinating:

- Life transitioned from uni to multicellular multiple times. Animals and land plants both evolved from single celled lineages independently. Fungi like Ascomycota and Basidiomycota went multicellular independently. Even green, red and brown algae went multicellular independently.

- Ctenophores (comb jellies) might have evolved nerves independently. Some sponges and placozoans have cells that perform coordinated, nerve‐like functions. These use a different genetic and cellular toolkit that synapses and seem to be an independent evolution.

- Octopi evolved brains independently of us. Our last common ancestor just had a primitive nerve net.

So at least the path from unicellular life to monkey/octopus level intelligence doesn't seem like a fluke at all to me. It seems like even on the same planet the major transitions occur repeatedly and independently.

The jump from monkey to human happened so fast that indicates to me its not vanishingly unlikely. (Monkeys as smart as chimps are thought to have evolved 15-20M years ago). You might say octopi and birds are older and haven't evolved but they didn't have hands to take advantage of their intelligence and create a positive feedback loop.

I'm trying to imagine where evolution and super-intelligence seem irreconcilable. Maybe it's the jump from monkeys to the Homo genus, where brain size first started to run away? I think its clear how increasing brain size can confer an advantage there once the stage is properly set. Chimps already had complex social dynamics and rudimentary tool use. We even see initial phases of modestly increasing brain size with the Australopiths that preceded Homo. However our hominid lineage is the only one that did have an increasing brain size.

Is it the progression from early Homo to us? That seems like a natural and inevitable progression once the process started.

Yes, devices to sense the environment and then to make some prediction about it seems quite complex to evolve by selection but to be expected given the tremendous advantage they can give. (And thank you for making the effort to illustrate the discussion with actual biology).

That's the evolution of the level of intelligence required to escape our environment that seems to miss the driver. I just can't see the evolutionary advantage. I would be the last person to underestimate animal intelligence, but none of the exemples, known and unknown, of brains that evolved on earth, have shown anything close to that.

I do acknowledge the fact that more individual as well as social intelligence is generaly good from a survival standpoint, but the gap in intelligence between lighting a fire or shaping a rock into a blade (both good for survival) into developing a theory of black holes or solving P=NP (both useless for survival) seems just too wide to be closed just by the hypothetical natural inertia in the brain growth process.

But maybe this is like emergence of intelligence in large neural network? That at some point, the more economical way for a neural network to be able to predict the correct output for a large set of input is to internalise some general knowledge and understanding of the laws governing the inputs? In other words, maybe the elements of reasoning and memory required to light a fire and carve rocks and understand seasons and prey behavior etc, are more easily synthesized into a brain that can, using the same elements of logic, develop a theory of black holes?

Maybe, but that take care only of the "individual" intelligence ; we still need to find a reason why social intelligence would go from "be smart enough to steal some bananas and therefore appear stronger and therefore can climb on more female at the end of the day" to "act and behave cooperatively toward a future common good". Is your solution to this the creation of super-organisms? If so, I'm no expert but I can think of as many exemples in nature of "super organisms" that became dumber (insect colonies) than more intelligent (multi-cellular) as a result.

I believe the root of my scepticism is actually my psssimistic views on social intelligence, the value of cooperation, that I doubt the central role of reason in most technical inovations, the role of faithful communication in the evolution of language, etc.

On another hand, you are the one who have to live with the Fermi paradox :)

Reply on news.ycombinator.com