[walleeee]

Ignoring the difficulty of defining such a needle in a general sense, I still think the point is disproved by the fact that groups can do things no individual member can conceive of, let alone do, alone.

If it's true of ants, who we likely agree have not so large a cognitive lightcone as us, surely it is of us. The only way I see for it not to be true of us is to posit a hard cap on intelligence which we are closer to.

There are also forms of intelligence which resist formalization and do not equate to cleverness. For instance, wisdom.

[IgorPartola]

I am specifically talking about intellectual capability. No one person can build an aircraft by themselves. But also if you take 10,000 people of IQ of exactly 100 (which is always defined as a median IQ for a population), you will not get a carrier either. You also wouldn’t get calculus or general relativity or genetic engineering.

[walleeee]

I'm not sure what to make of the point about 1 or 10,000 people. 1 person can definitely build a primitive aircraft with time, money and perseverance. No 10,000 random people, whether IQ 100 or 1000, will build an airliner, but Airbus will, more or less reliably regardless of the exact IQ of any particular employee. The ability to build one is distributed over hundreds of thousands of people and millions of machines at Airbus, and millions of people and billions of machines in the industrial civilization behind the whole endeavor.

If your point is just that different people have different capacities, then sure, that's true enough. But even so, it's an extremely coarse summary of an enormously complex and detailed landscape.

[IgorPartola]

I am talking about building one of these: https://cdn.britannica.com/34/94834-050-16DB7EEB/USS-Carl-Vi...

I think I am not being clear in what I am saying. My point is that if you have a group of 10k average individuals their collective intelligence will not let them design something like this even though they will have the manpower and presumably the raw materials to build it. But if you add a small handful of much smarter than average individuals with expertise in the relevant engineering fields and they would not only be able to conceive of it as a concept but also design and it and lead the larger group to building it.

This does not absolve the group of needing to recognize those who are the most intellectually capable amongst them and put them into a leadership role. But the opposite is true because you cannot by definition imagine someone who is significantly smarter than you and you cannot emulate it. Collective intelligence of a group is bounded by the smartest person there. Adding more average intelligence to a group will not make it smarter, just larger, and potentially get in the way of the group recognizing the intelligence of its smartest members.

[walleeee]

If your point is that competent people must direct the activities of a group in order to accomplish a complicated task, sure. That's almost always true.

> Collective intelligence of a group is bounded by the smartest person there.

I maintain this is false. Who was the smartest person in the Vienna circle? A group of competent, well aligned people is potentially a much more capable thing than the one member with the highest IQ score alone. Collective intelligence is no simple combination of the capacities of its constituents. That is what the anthill demonstrates. The capacities of individuals do limit the capacities of the whole in some way, but it is nowhere near as simple as "bounded by the smartest person". This has a lot to do with the fact that "smartest person in the room" is not a particularly meaningful title absent a precise definition of what intelligence/competence means in your context of interest.

I agree with your point that adding people to an organization does not necessarily improve its capacities. Collective intelligence is not simply additive, nor is it simply bounded by the best individual on any particular dimension

[IgorPartola]

Ok let me reframe it like this: the collective intelligence of a random group of people is bounded by the collective intelligence of its top N% members, ranked by intelligence, where N is some relatively small number.”

If you assume a bell curve for intelligence, then it is primarily the people who are two standard deviations and up from median that are driving innovation, assuming the group allows such individuals to do so. In a small group this might be just one person, in a nation of hundreds of millions it would be single digits of millions, etc.

If the whole group is actively hostile towards this or there is a systematic barrier to it, such as the intelligentsia purge under Stalin, then there is probably going to be another factor that is much stronger at bounding the maximum intelligence.

I postulates that this holds true for any group where the average intelligence is close to 100 IQ or any other metric you want to use to designate the median intelligence in the human population overall.

If you get together a group of people who are already all over two standard deviations up from the overall median you will probably get a different calculation for the upper bound because higher order terms will become more significant, such as specialities, age, experience, etc. But when we are talking about societies, you are always talking about comparing to the median. As far as I know something like a median intelligence slowly trends up in humans over centuries but it is slow enough that there aren’t societies where their median intelligence is a lot lower or higher than any other society in existence at the same time.

Ants are a terrible example due to using a completely different intelligence model. They are incapable of recognizing innovation, so even if you throw in a group of Fermat ants into the mix, they will still end up working as just slightly more efficient drones. This fits into the model above as the society being set up with an active barrier to allowing its smartest members to innovate.

[walleeee]

> [ants] are incapable of recognizing innovation, so even if you throw in a group of Fermat ants into the mix, they will still end up working as just slightly more efficient drones.

And yet the anthill does very complex things even though its Fermats might feel their unique talents go unused. This strengthens the case unless you posit that humankind is so close to a universal intelligence limit that we can see its precise contours and can therefore know exactly what can be known and done, and how, exhaustively.

Separately, everything you say needs an operationalization of intelligence. I understand IQ is your proxy of choice. Presumably this measure correlates to innovation in such things as physics and mathematics. Presumably also to innovation in speculative financial vehicles, elaborate social schemes, weapons of mass destruction, and success in the war against nature. Will you adjust your definition when some of its manifestations come into tension with others, or stick with it come what may?

There is no intelligence absent context. Intelligence is a relation between an agent and everything else relevant, with respect to a goal.