How Smart Are Humans?

My stock counterargument to this: insofar as humans' advantage over other animals stems primarily from our ability to transmit knowledge/memes/etc across individuals and through generations, we should expect AI to have a much larger advantage, because they can do the same thing far, far better. This doesn't even require the AI to be all that "smart" - even just the ability to copy minds directly would allow transmission from "parent" to "child" with far less knowledge-loss than humans can achieve. (Consider, for instance, the notorious difficulty of training a student to the same level as an unusually accomplished human - and then imagine how much more effective it would be to just copy the accomplished human.) And that's just copying. The ability to merge two minds does not necessarily come for free with any AI (as copying does), but will likely be far easier for AI than for humans, and would provide a much more extreme version of what humans achieve by passing information via language. (Consider, for instance, the notorious difficulty of communicating well and avoiding misunderstanding - and then imagine how much easier that would be if two minds or components-of-minds could merge.)

What these arguments suggest is that humans' language capabilities are themselves relatively primitive ways to transmit knowledge/memes/etc between individuals, and it is not in-principle hard to do better once one starts to optimize directly for that goal without the constraints of biological wetware. More generally, we should expect that whatever did give humans such a large advantage (be it language or something else), humans have a relatively primitive version of it, because otherwise humans would have taken off sooner in our evolutionary history. As the saying goes, "humans are as stupid as they can be while still taking over the world, because otherwise they would have done so earlier".

It is important to remember that humans, unlike all other species, are able to use complex language. This is a huge confounding factor, when we try to compare the intelligence of humans and animals. It is obviously very powerful to be able to exchange complex ideas, and build up knowledge intergenerationally. This would probably be enough to give humans a very large advantage, even if our intelligence was otherwise exactly the same as that of other primates.

Communication is an aspect of intelligence. It takes place in the brain not the kidneys. Now you could argue that communication is a special extra boost above and beyond the normal gains of intelligence, that humans are near the top of the communication sigmoid, and that there are no other special extra boosts out there. 

Do monkeys have a mind capable of understanding calculus internally and just lack any language capable of learning it? (Such that a monkey given perfect communication but no other increases in intelligence would be able to learn calculus.) 

I am not convinced the question is meaningful. I doubt that "communication" is a clear boundary on the neurochemical level, with a sharp divide between communication neurons and other neurons.

First of all, there have been "feral" humans that grew up surrounded by animals. As far as I know, these humans are not obviously much more intelligent than animals (in terms of their ability to solve problems).

Think of that like a modern supercomputer being used to play pong. 

(well not that big of a gap, but you get the picture)

Animal brains have a relatively simple and limited range of pieces of software they can run. 

Human brains are able to run a much wider range of much more complicated programs. 

In other words, human intelligence shows up in that with the right training, we are able to do all sorts of complicated things. Whereas there are plenty of things most humans can do that animals can't be trained to do. 

To actually be able to do useful stuff, humans need the right training, both in the specific technical details, and more general stuff like the scientific method. (with some ability to figure those things out from large amounts of trial and error)

Train a human in a nutty cult, and their intelligence is useless. But the point is that humans can be trained to do physics. Not that every human crawls out of the womb doing calculus. 

I agree that if your only examples of humans were feral humans, then you would have no reason to think humans were much smarter. And then you would be very surprised by basically any educated human. 

Or another way to put this, for a slightly different definition of the word "intelligence" is that humans can be much more intelligent than animals with the right environment.

I agree with your thesis that language is critical to human intelligence.

I'm curious if you have specific references for the claim that humans raised without language are far less smart. I've tried to find the best evidence, and the only decently documented cases I could find might have been developmentally disabled. The caretakers of severely neglected children (e.g., no one talks to them) universally claim that those children were nonverbal, and that's why they were treated as nonhuman. Stories of literal wolf children are much less well documented.

This is an academic question; I'd like to have better evidence. I strongly suspect that a developmentally normal human would be vastly less intelligent if raised without language. I'd like to have stronger evidence for this claim.

However, if it is the case that the difference between humans and monkeys is mostly due to a one-shot discrete difference (ie language), then this cannot necessarily be repeated to get a similar gain in intelligence a second time.

Perhaps language is a zero-one, i.e. language renders a mind "cognitively complete" in the sense that the mind can represent anything about the external world, and make any inferences using those representations. But intelligence is not thereby zero-one because intelligence depends on continuous variables like computional speed, memory, etc.

More concretely, I am sceptic that "we end up with AI geniuses, but not AI gods", because running a genius at 10,000x speed, parallelised over 10,000x cores, with instantaneous access to the internet does (I think) make an AI god. A difference is quantity is a difference in kind.

Thar said, there might exist plausible threat models which require an AI which doesn't spatiotemporally decompose into less smart AIs. Could you sketch one out?

humans have an innate ability to use complex, combinatorial language

The important part is that we use this language not just when talking to each other, but also internally. We can remember spoken sentences (for example other people's commands). We can talk to ourselves (i.e. think verbally). We can talk to our future selves (e.g. by writing a to-do list). We can communicate asynchronously (i.e. write). We can follow complex written instructions.

I am writing it this way to express that merely saying "language" feels like understatement of all the abilities that are unlocked by non-standard uses of language.

What if: Human ability to use language with certain interesting and useful properties isn't particularly innate, though we may have evolved advantages with it over time. Rather, it is a cultural invention with useful applications to technology. If this idea is true we should see gains in capability by any animal society that culturally transmits traditions that learns or develops languages of similar properties. I think from the AI perspective we can certainly argue it's been quite handy for a nonhuman entity that does have innate connection to language.

At the moment, the most common answers to these questions are “much more intelligent”, and “very quickly”.

For "very quickly" ("foom", intelligence explosion), the aspect which is particularly important is how smart humans are specifically with respect to the task of making better and better AI systems.

We strongly suspect that AIs which are "roughly human equivalent with respect to the task of making better and better AI systems" will rapidly become much smarter than humans specifically with respect to this task, and this is the commonly assumed mechanism of "hard take-off'.

Other kinds of smartness are expected to be mostly downstream of this central mechanism...

(So, in some sense, when we are talking about take-off speed, we are talking not so much about "general AIs", but about systems which are strong "artificial AI researchers and software engineers").