I like the endpoint, but I have many thoughts on the data and logic used to get there (as you might expect from a mechanical engineer). For one, it is extremely common in hardtech that engineers model how something works before physicists define a comprehensive theory or law for why it works that way. AlphaFold's ability to predict protein folding means there must be an underlying logic, even if we don't understand it yet. Shortcuts like that are great! Engineers usually only care that it works, but the lack of understanding can limit applications or introduce failure modes we don't really understand. The humanities would label this situation as, "Our reach exceeds our grasp." That is both warning and invitation.
My understanding of the defining characteristic of modernity versus "post-modernism" and its derivatives is belief. Modernity was centered on belief, post-modernism saw belief as dangerous and tried to hide it or convince people to abandon beliefs because of its contradictions (see how Ayn Rand is treated). But of course we need belief—in ourselves, in our families, our communities, the future. So it all crept back in, including malevolent beliefs. If the future you imagine is fragile, I have to wonder if it is fragile because of a belief or a lack of one.
Which brings us back to the family, as the core of an anti-fragile future. That is the truth, probably tautologically so. I share the belief that the purpose of all this technology, AI, spaceships, computer chips, is to embrace and deepen our own humanity. I'm not sure that's at odds with specialization, so long as the specialization is a choice of our own and not just a set of incentives laid upon us by society that gives us certainty of survival but not much else. I suspect there are other anti-fragile virtues to be included as part of that future, at the center of which is resilience. In The Ballad of the White Horse, Chesterton argues that it is not our purpose to give in to either despair or presume victory. Like the little flap that could, our purpose is to get up time and again to meet the challenges of today and tomorrow:
> AlphaFold's ability to predict protein folding means there must be an underlying logic, even if we don't understand it yet.
I agree that the fact that a deep learning model can predict protein folding means there are patterns, but that doesn't mean there needs to be an "underlying logic" in the sense of a compact set of clearly-defined principles that tell you what structure a given protein will have. My understanding of how these models work is that they learn a gigantic muddle of partial rules, exceptions to those rules, exceptions to the exceptions, ad effectively infinitum. The "principles" that AlphaFold has learned may be literally millions of separate ideas such as "when these 13 sub-patterns all appear in the same general region of a protein, there is an 11% higher probability that the ninth of those sub-patterns will fold so as to bring it into conjunction with the third sub-pattern".
(And often these principles will not be universal truths, but merely things that tend to be true in practice in the context of proteins that appear in nature – just as chess grandmasters are very good at memorizing real chess positions, but not at memorizing random arrangements of pieces.)
I don't think Newton had a pocket watch. Harrison didn't produce the compact marine chronometer to solve the longitude problem until about 1740, but Newton died in 1727.
Talking about western fertility's decline, and attributing it to Western intellectual trends, seems strange when Bangladesh, South Korea, and Pakistan show substantial declines and Israel and New Zealand are outperformers (though still fairly low). Do you think Bangladeshi fertility collapse is a coincidence?
I do sense a bit of irony here - the criticism of modernity you put forth strikes me as analogous to common post-modernist thought: Especially the skepticism of reductionism - calling for viewing concepts not in isolation but as only being understandable through its relation with their systems. Likewise "embracing mystery" rather than demanding rational explanation is particularly postmodernist.
My understanding is that where you diverge is that you do *not* actually propose departing from enlightenment rationalism/ ordering nature and modernism. You're rather saying our current tools have hit walls in solving this classic modernist problem and we need complex systems thinking + the associated tech (AI) to do so - but at the core this doesn't strike me as a rejection of "nature as a machine"?
BTW - I am not contrasting your vision to post-modernist visions becaus a) my understanding of post-modernism is limited b) my limited understanding is that such a notion would be rejected by those that believe the need to 'control nature as a machine' would have to be rejected and c) I (therefore) haven't really seen great alternative visions of the future yet that would be 'post-modernist' beyond 'we have to criticise power structures, understand history + context for suboptimal situations and move back to smaller units of society [which most post-modernists, as I understand it, wouldn't (have to? shouldn't?) be a family though...]'
I like the endpoint, but I have many thoughts on the data and logic used to get there (as you might expect from a mechanical engineer). For one, it is extremely common in hardtech that engineers model how something works before physicists define a comprehensive theory or law for why it works that way. AlphaFold's ability to predict protein folding means there must be an underlying logic, even if we don't understand it yet. Shortcuts like that are great! Engineers usually only care that it works, but the lack of understanding can limit applications or introduce failure modes we don't really understand. The humanities would label this situation as, "Our reach exceeds our grasp." That is both warning and invitation.
My understanding of the defining characteristic of modernity versus "post-modernism" and its derivatives is belief. Modernity was centered on belief, post-modernism saw belief as dangerous and tried to hide it or convince people to abandon beliefs because of its contradictions (see how Ayn Rand is treated). But of course we need belief—in ourselves, in our families, our communities, the future. So it all crept back in, including malevolent beliefs. If the future you imagine is fragile, I have to wonder if it is fragile because of a belief or a lack of one.
Which brings us back to the family, as the core of an anti-fragile future. That is the truth, probably tautologically so. I share the belief that the purpose of all this technology, AI, spaceships, computer chips, is to embrace and deepen our own humanity. I'm not sure that's at odds with specialization, so long as the specialization is a choice of our own and not just a set of incentives laid upon us by society that gives us certainty of survival but not much else. I suspect there are other anti-fragile virtues to be included as part of that future, at the center of which is resilience. In The Ballad of the White Horse, Chesterton argues that it is not our purpose to give in to either despair or presume victory. Like the little flap that could, our purpose is to get up time and again to meet the challenges of today and tomorrow:
Night shall be thrice night over you,
And heaven an iron cope.
Do you have joy without a cause,
Yea, faith without a hope?
> AlphaFold's ability to predict protein folding means there must be an underlying logic, even if we don't understand it yet.
I agree that the fact that a deep learning model can predict protein folding means there are patterns, but that doesn't mean there needs to be an "underlying logic" in the sense of a compact set of clearly-defined principles that tell you what structure a given protein will have. My understanding of how these models work is that they learn a gigantic muddle of partial rules, exceptions to those rules, exceptions to the exceptions, ad effectively infinitum. The "principles" that AlphaFold has learned may be literally millions of separate ideas such as "when these 13 sub-patterns all appear in the same general region of a protein, there is an 11% higher probability that the ninth of those sub-patterns will fold so as to bring it into conjunction with the third sub-pattern".
(And often these principles will not be universal truths, but merely things that tend to be true in practice in the context of proteins that appear in nature – just as chess grandmasters are very good at memorizing real chess positions, but not at memorizing random arrangements of pieces.)
I don't think Newton had a pocket watch. Harrison didn't produce the compact marine chronometer to solve the longitude problem until about 1740, but Newton died in 1727.
Talking about western fertility's decline, and attributing it to Western intellectual trends, seems strange when Bangladesh, South Korea, and Pakistan show substantial declines and Israel and New Zealand are outperformers (though still fairly low). Do you think Bangladeshi fertility collapse is a coincidence?
I do sense a bit of irony here - the criticism of modernity you put forth strikes me as analogous to common post-modernist thought: Especially the skepticism of reductionism - calling for viewing concepts not in isolation but as only being understandable through its relation with their systems. Likewise "embracing mystery" rather than demanding rational explanation is particularly postmodernist.
My understanding is that where you diverge is that you do *not* actually propose departing from enlightenment rationalism/ ordering nature and modernism. You're rather saying our current tools have hit walls in solving this classic modernist problem and we need complex systems thinking + the associated tech (AI) to do so - but at the core this doesn't strike me as a rejection of "nature as a machine"?
BTW - I am not contrasting your vision to post-modernist visions becaus a) my understanding of post-modernism is limited b) my limited understanding is that such a notion would be rejected by those that believe the need to 'control nature as a machine' would have to be rejected and c) I (therefore) haven't really seen great alternative visions of the future yet that would be 'post-modernist' beyond 'we have to criticise power structures, understand history + context for suboptimal situations and move back to smaller units of society [which most post-modernists, as I understand it, wouldn't (have to? shouldn't?) be a family though...]'