Hmmm this one is definitely underspecified, but I suppose that's what makes it interesting to ponder. Just considering scientific knowledge, humanity has clearly developed some algorithms that lead us to truth (namely, the scientific method). We may also include dumb trial and error, whatever scientists do in their heads to think of new experiments/hypotheses, and the norms and values of the scientific community. So maybe we can think of the scientific community as an instantiation of this truth algorithm and Bach is asking what is the best possible way to arrange a scientific community. Bach is an AI researcher so I suppose he would then wonder if we could use this algorithm to create an AI scientific community that automates the generation of scientific truth.
And here I am, courtesy of Scott Alexander. I've been a fan of the Edge questions for many years. Here's some semi-final answers.
Why is Homo sapiens the sole non-extinct species of hominin? Because the other species interbred to create us, and we out-competed them.
Will the process of discovery be completed in any of the natural sciences? Theoretically yes, but probably not.
Will some things about life, consciousness, and society necessarily remain unseen? Yes.
Are there limits to what we can know about the universe? Yes.
Is it possible to control a system capable of evolving? I doubt it.
What would the ability to synthesize creativity do to cultural evolution? I don't think it's possible to synthesize creativity.
Will human psychology keep pace with the exponential growth of technological innovation associated with cultural evolution? The last few decades have shown that we can't keep pace.
Can technology tame evolution? I doubt it.
How far are we from wishing to return to the technologies of the year 1900? Extremely far.
Are complex biological neural systems fundamentally unpredictable? Yes.
Is a human brain capable of understanding a human brain? No.
Is there a way for humans to directly experience what it’s like to be another entity? No.
Will a machine ever be able to feel what an organism feels? No.
Can we acquire complete access to our unconscious minds? No.
How many incommensurable ideas can we hold in our mind simultaneously? I can do 2.
How can we reap the benefits of the wide and open exchange of data without undermining the values that depend upon the scarcity of information? There are values that depend upon the scarcity of information?
Are people who cheat vital to driving progress in human societies? I doubt it.
Can we design a modern society without money which is at least as effective economically and politically as our current system? Yes, but I don't know what it would be.
Can we re-design our education system based on the principle of neurodiversity? We could, but will we?
Why should we prize the original object over a perfect replica? We shouldn't, but people with the collector mentality do.
Are stories bad for us? No, they are essential for social beings.
Does something unprecedented happen when we finally learn our own source code? Yes, but I doubt we will.
When in the evolution of animal life did the capacity to experience love for another being first emerge? Depends of the definition of love. For one reasonable literary definition, it was the Troubadour era, 12th century.
Can natural selection's legacy of sex differences in values be reconciled with the universal values of the Enlightenment? No.
Will scientific advances about the causes of sexual conflict help to end the "battle of the sexes"? They can help, but won't completely eliminate it.
What will happen to religion on earth when the first alien life form is found? Depends on the specific religion. Some would not be affected.
How much would surrendering our god(s) strengthen the odds of our survival? Maybe zero.
"What is the optimal algorithm for discovering truth?
— Joscha Bach"
How would we even recognize truth, much less discover it?
Hmmm this one is definitely underspecified, but I suppose that's what makes it interesting to ponder. Just considering scientific knowledge, humanity has clearly developed some algorithms that lead us to truth (namely, the scientific method). We may also include dumb trial and error, whatever scientists do in their heads to think of new experiments/hypotheses, and the norms and values of the scientific community. So maybe we can think of the scientific community as an instantiation of this truth algorithm and Bach is asking what is the best possible way to arrange a scientific community. Bach is an AI researcher so I suppose he would then wonder if we could use this algorithm to create an AI scientific community that automates the generation of scientific truth.
incredible nerdbait here
Oh definitely... come hither little nerdles...
And here I am, courtesy of Scott Alexander. I've been a fan of the Edge questions for many years. Here's some semi-final answers.
Why is Homo sapiens the sole non-extinct species of hominin? Because the other species interbred to create us, and we out-competed them.
Will the process of discovery be completed in any of the natural sciences? Theoretically yes, but probably not.
Will some things about life, consciousness, and society necessarily remain unseen? Yes.
Are there limits to what we can know about the universe? Yes.
Is it possible to control a system capable of evolving? I doubt it.
What would the ability to synthesize creativity do to cultural evolution? I don't think it's possible to synthesize creativity.
Will human psychology keep pace with the exponential growth of technological innovation associated with cultural evolution? The last few decades have shown that we can't keep pace.
Can technology tame evolution? I doubt it.
How far are we from wishing to return to the technologies of the year 1900? Extremely far.
Are complex biological neural systems fundamentally unpredictable? Yes.
Is a human brain capable of understanding a human brain? No.
Is there a way for humans to directly experience what it’s like to be another entity? No.
Will a machine ever be able to feel what an organism feels? No.
Can we acquire complete access to our unconscious minds? No.
How many incommensurable ideas can we hold in our mind simultaneously? I can do 2.
How can we reap the benefits of the wide and open exchange of data without undermining the values that depend upon the scarcity of information? There are values that depend upon the scarcity of information?
Are people who cheat vital to driving progress in human societies? I doubt it.
Can we design a modern society without money which is at least as effective economically and politically as our current system? Yes, but I don't know what it would be.
Can we re-design our education system based on the principle of neurodiversity? We could, but will we?
Why should we prize the original object over a perfect replica? We shouldn't, but people with the collector mentality do.
Are stories bad for us? No, they are essential for social beings.
Does something unprecedented happen when we finally learn our own source code? Yes, but I doubt we will.
When in the evolution of animal life did the capacity to experience love for another being first emerge? Depends of the definition of love. For one reasonable literary definition, it was the Troubadour era, 12th century.
Can natural selection's legacy of sex differences in values be reconciled with the universal values of the Enlightenment? No.
Will scientific advances about the causes of sexual conflict help to end the "battle of the sexes"? They can help, but won't completely eliminate it.
What will happen to religion on earth when the first alien life form is found? Depends on the specific religion. Some would not be affected.
How much would surrendering our god(s) strengthen the odds of our survival? Maybe zero.
I remember being obsessed with Edge in high school...no idea how I originally discovered it. thanks for the throwback