Question
50 people to run the world: Who are they and why?
Answer
Note: I am ignoring the "rules" because they make this an impossible problem.
I would choose 40 people via random sampling among people above the age of 4, with perhaps some extra weight for the 14+ crowd, and 10 randomly from a pool of people who could be said to have actually run significant portions of the world before (probably current leaders of sufficiently large nations, corporations etc... the pool might be "anyone who has at least nominally led more than 100 million people?).
The former because any more deliberate strategy will inevitably build in bias for specific mental models, and the latter because "running the world" actually takes some skill and experience, believe it or not, so no matter how much you might hate the people doing it right now, they're the only ones who have meaningful things to say about how to do it and share direct experiences. I wouldn't trust a group that does not include at least a few people who've done something like it before.
Then I'd let the 50 figure out the rest. This model, I believe, would be more democratic than democracy itself. The world is now clearly too big and complex for democracy, but older systems are even worse. We need innovation. Random representation is the last frontier. And yeah, every 4 years, I'd swap out half the 40+10. Where the 10 would not be the original 10 (who would obviously naturally lead-by-authority to some extent), but the emergent 10. Randomly axe half of them, irrespective of demonstrated effectiveness.
This still doesn't address the tyranny of the majority problem (i.e. tyranny of small random samples, which will inevitably exclude 90% of the small populations that are identified primarily by long-tail variables), but given the constraints of the thought experiment, any attempt to address "tyranny of the majority" will be a cure that's worse than the disease.
Basically, 50 is simply too small for a planet of 7 billion, with many times that many voiceless stakeholders (farmed animals etc.) I'd say the thought experiment would need like 5000 people, along with significantly more structure (because it is like 30 times the Dunbar number and will not self-organize effectively...any group above 150 will almost certainly self-organize along identity politics lines and ruin the whole experiment).
So my counter-proposal would be 5000 people, also randomly selected (with some effort to intelligently include outliers that would be missed even at this scale), and randomly internally organized into 150-person cells, with some room for internal social mobility.
In a way, the test for this alternate thought experiment is self-organization. If they cannot even self-organize to become an effective governance body, they certainly cannot run the world.
But 50 is almost without a doubt waaaaay too small. I think people really don't get how big 7 billion is.
7000000000.
Organized in 250+ nations currently.
Speaking at least 100 major languages with more than some X million speakers each.
In a world with 100s of large companies, and MILLIONS of smaller ones.
And don't forget gazillions of chickens and cattle and other animals currently being tortured in various parts of the planet.
This is a serious answer, not whimsical or thought-experimentish. I truly believe in the power of randomness and the idea that it is the only solution beyond a certain level of complexity. We already know from developments in computer science and algorithms in the last 30-40 years just how hard high-dimensionality problems are, and how effective randomness (properly applied) is, as a solution.
Governing the world today, no matter how you define the problem, is the equivalent of a super EXPSPACE-hard set of conflicting problems. Social scientists have started calling these problems "wicked", but that's a weak human-moral adjective for the sheer complexity of the world we're in today.
We need to give up our attachment to democracy as the ultimate ideal, our attachment to apparently sound ideas of judging success/failure and causation patterns. We need to give up our attachment to "meritocracy" (at best a very false sense of security, because "merit" itself is one of the intractable problems that must be continuously solved). And most importantly, we must give up our attachment to "values" of various sorts, no matter how compelling, and take the plunge into the sort of randomness that might (say) put an Idi Amin into the 50 (or 5000). Without taking that huge downside risk, we also will not be able to avail ourselves of huge upside opportunities. Instead, we'll end up with an interchangeable parade of impotent Obamas and Romneys who reassure and pacify but cannot actually break out of what is essentially a global governance gridlock. Because they are part of a determinstic governance algorithm for a system that has outgrown the capacity of deterministic systems and requires serious, deliberate injection of true randomness.
If you want a good random-number generator that is not pseudorandom btw, Hojun Song's "open source satellite" will soon be generating a random number signal from cosmic background radiation for all to use. A "random positioning system" if you will.
I would choose 40 people via random sampling among people above the age of 4, with perhaps some extra weight for the 14+ crowd, and 10 randomly from a pool of people who could be said to have actually run significant portions of the world before (probably current leaders of sufficiently large nations, corporations etc... the pool might be "anyone who has at least nominally led more than 100 million people?).
The former because any more deliberate strategy will inevitably build in bias for specific mental models, and the latter because "running the world" actually takes some skill and experience, believe it or not, so no matter how much you might hate the people doing it right now, they're the only ones who have meaningful things to say about how to do it and share direct experiences. I wouldn't trust a group that does not include at least a few people who've done something like it before.
Then I'd let the 50 figure out the rest. This model, I believe, would be more democratic than democracy itself. The world is now clearly too big and complex for democracy, but older systems are even worse. We need innovation. Random representation is the last frontier. And yeah, every 4 years, I'd swap out half the 40+10. Where the 10 would not be the original 10 (who would obviously naturally lead-by-authority to some extent), but the emergent 10. Randomly axe half of them, irrespective of demonstrated effectiveness.
This still doesn't address the tyranny of the majority problem (i.e. tyranny of small random samples, which will inevitably exclude 90% of the small populations that are identified primarily by long-tail variables), but given the constraints of the thought experiment, any attempt to address "tyranny of the majority" will be a cure that's worse than the disease.
Basically, 50 is simply too small for a planet of 7 billion, with many times that many voiceless stakeholders (farmed animals etc.) I'd say the thought experiment would need like 5000 people, along with significantly more structure (because it is like 30 times the Dunbar number and will not self-organize effectively...any group above 150 will almost certainly self-organize along identity politics lines and ruin the whole experiment).
So my counter-proposal would be 5000 people, also randomly selected (with some effort to intelligently include outliers that would be missed even at this scale), and randomly internally organized into 150-person cells, with some room for internal social mobility.
In a way, the test for this alternate thought experiment is self-organization. If they cannot even self-organize to become an effective governance body, they certainly cannot run the world.
But 50 is almost without a doubt waaaaay too small. I think people really don't get how big 7 billion is.
7000000000.
Organized in 250+ nations currently.
Speaking at least 100 major languages with more than some X million speakers each.
In a world with 100s of large companies, and MILLIONS of smaller ones.
And don't forget gazillions of chickens and cattle and other animals currently being tortured in various parts of the planet.
This is a serious answer, not whimsical or thought-experimentish. I truly believe in the power of randomness and the idea that it is the only solution beyond a certain level of complexity. We already know from developments in computer science and algorithms in the last 30-40 years just how hard high-dimensionality problems are, and how effective randomness (properly applied) is, as a solution.
Governing the world today, no matter how you define the problem, is the equivalent of a super EXPSPACE-hard set of conflicting problems. Social scientists have started calling these problems "wicked", but that's a weak human-moral adjective for the sheer complexity of the world we're in today.
We need to give up our attachment to democracy as the ultimate ideal, our attachment to apparently sound ideas of judging success/failure and causation patterns. We need to give up our attachment to "meritocracy" (at best a very false sense of security, because "merit" itself is one of the intractable problems that must be continuously solved). And most importantly, we must give up our attachment to "values" of various sorts, no matter how compelling, and take the plunge into the sort of randomness that might (say) put an Idi Amin into the 50 (or 5000). Without taking that huge downside risk, we also will not be able to avail ourselves of huge upside opportunities. Instead, we'll end up with an interchangeable parade of impotent Obamas and Romneys who reassure and pacify but cannot actually break out of what is essentially a global governance gridlock. Because they are part of a determinstic governance algorithm for a system that has outgrown the capacity of deterministic systems and requires serious, deliberate injection of true randomness.
If you want a good random-number generator that is not pseudorandom btw, Hojun Song's "open source satellite" will soon be generating a random number signal from cosmic background radiation for all to use. A "random positioning system" if you will.