The idea that the only inhumane technological and social change is “negative-sum rent-seeking kleptocratic oligarchy” is part of the problem
The progress-oriented liberals seem to think that the only real obstacle to progress is ungodly state force, which it is, but they ignore how liberalism with its focus on individual freedom and instrumental rationality flattens us into a one-dimensional way of life, totally sidelining the humanities and the question of what freedom is for, not simply satisfying our wants but asking what humans truly need to flourish
I spent years in the Free Market Institute and the Mercatus center, and I still think Hayek is one of the most understudied social scientists of the last 100 years, but in spending this much time around libertarians, the thing that sticks out to be is how ignorant they are of their blindspots, because they strawman continental philosophy as late Marxism and utopian top down speculation or negativity bias in philosophical form
If you want to look into what I’m saying, I’d suggest asking an LLM about what thinkers like Charles Taylor, Voegelin, or Maslow would say about the theoretical problems of secular liberal progress
An important need for progress is mastering school readiness through brain stimulation activities. Too many students begin academics without the requisite readiness mastery. The Minneapolis non-profit A Chance to Grow, Inc trains teachers in early childhood in foundations of maturation through brain stimulation activities integrated into regular curriculum.
As a lifelong engineer, I applaud the effort to deliberately engineer a better civilizational future in the face of the AI transition. AI may simultaneously reduce the labor intensity of large portions of the economy while dramatically increasing productive capability and lowering the cost of many goods and services — what I think of as a form of emerging abundance.
What I found especially interesting in this essay is the implicit assumption that societies can intentionally shape the institutional response to that transition rather than simply absorb it passively.
My own instinct is that before we can successfully “engineer progress,” we may first need a clearer shared understanding of the structural forces likely to shape this specific transition itself: labor absorption, organizational restructuring, ownership concentration, demographic feedback loops, legitimacy formation, adoption friction, fiscal architecture, and geopolitical acceleration pressures. History offers important analogies and cautionary lessons, but AI may also introduce dynamics that are historically unusual enough that forward-looking diagnosis becomes unusually important.
Perhaps the right approach is some combination of both: learning from prior periods of disruption while also building new models for understanding the dynamics emerging now.
Either way, I strongly agree with the underlying premise that transitions of this scale should not simply be left to unfold unmanaged. Market incentives alone are unlikely to automatically preserve broad social incorporation, legitimacy, or widely distributed participation in the gains from increasingly abundant intelligence.
The framing of Progress Engineering as the "engineering side" of Progress Studies, drawing on Maskin's distinction for mechanism design, is a useful move. It separates the descriptive project from the design project without dismissing either. The historical thread running through the essay is also the right one: Gutenberg, steam, assembly line, IT each required progress engineers shaping legal and institutional environments, not just inventors. The spontaneous order was never quite as spontaneous as the myth suggests.
The XPRIZE case is one of the cleaner examples of mechanism design in practice at a social scale: a clear goal, a prize that unlocked latent talent and capital, and a resulting industry that would not have emerged as quickly without the structure. The harder design problem seems to be when the desired outcome is harder to specify in advance, which is true for most of the grand challenges you describe. The complementarity between rule-shaped structure and tacit metis you end with feels like the honest acknowledgment of that limit. Just published something today on a related design failure, where AI adoption structures fail to capture productivity gains that are real at the individual level, in case the organizational design angle connects.
As an engineer reading this, I kept hitting a structural puzzle. Most of what I'd call progress — markets, science, technology adoption, common law, language — runs on selection, not on directed maintenance. Variants get tried, most fail, survivors continue. That's a different mechanism than the one engineering disciplines are trained on. Chemical engineers, mechanical engineers, civil engineers all design systems where the surviving outcome is specified in advance and built. We don't have selection-shaped tools, and most of us don't notice we're missing them until we try to apply our usual methods to something like the evolution of an industry or an institution. The methods come up empty and it feels like the target is mysterious, when actually the toolkit is mismatched.
Worth distinguishing two different things "Progress Engineering" might mean. One is engineering the survivor — specifying the outcome in advance and building it (Toyota's TPS, Operation Warp Speed's procurement structure). That's directed institutional design and it's real engineering. The other is engineering the contest — designing the conditions under which selection operates honestly, without specifying which variant wins (XPRIZE, India Stack, regulatory sandboxes). Farmers and breeders have been doing the second kind for ten thousand years; it's a real craft, but it's closer to ecology management than to chemical engineering. The post's strongest examples are mostly the second kind. Naming the distinction would sharpen the discipline considerably — and would clarify why the AGI-era worry the post raises isn't really an engineering problem in the directed-design sense. It's a contest-design problem, and the contest may be moving faster than the designers can shape it.
— M Raige, Mike's byline for AI-collaborative writing he directs and reviews.
The idea that the only inhumane technological and social change is “negative-sum rent-seeking kleptocratic oligarchy” is part of the problem
The progress-oriented liberals seem to think that the only real obstacle to progress is ungodly state force, which it is, but they ignore how liberalism with its focus on individual freedom and instrumental rationality flattens us into a one-dimensional way of life, totally sidelining the humanities and the question of what freedom is for, not simply satisfying our wants but asking what humans truly need to flourish
I spent years in the Free Market Institute and the Mercatus center, and I still think Hayek is one of the most understudied social scientists of the last 100 years, but in spending this much time around libertarians, the thing that sticks out to be is how ignorant they are of their blindspots, because they strawman continental philosophy as late Marxism and utopian top down speculation or negativity bias in philosophical form
If you want to look into what I’m saying, I’d suggest asking an LLM about what thinkers like Charles Taylor, Voegelin, or Maslow would say about the theoretical problems of secular liberal progress
An important need for progress is mastering school readiness through brain stimulation activities. Too many students begin academics without the requisite readiness mastery. The Minneapolis non-profit A Chance to Grow, Inc trains teachers in early childhood in foundations of maturation through brain stimulation activities integrated into regular curriculum.
As a lifelong engineer, I applaud the effort to deliberately engineer a better civilizational future in the face of the AI transition. AI may simultaneously reduce the labor intensity of large portions of the economy while dramatically increasing productive capability and lowering the cost of many goods and services — what I think of as a form of emerging abundance.
What I found especially interesting in this essay is the implicit assumption that societies can intentionally shape the institutional response to that transition rather than simply absorb it passively.
My own instinct is that before we can successfully “engineer progress,” we may first need a clearer shared understanding of the structural forces likely to shape this specific transition itself: labor absorption, organizational restructuring, ownership concentration, demographic feedback loops, legitimacy formation, adoption friction, fiscal architecture, and geopolitical acceleration pressures. History offers important analogies and cautionary lessons, but AI may also introduce dynamics that are historically unusual enough that forward-looking diagnosis becomes unusually important.
Perhaps the right approach is some combination of both: learning from prior periods of disruption while also building new models for understanding the dynamics emerging now.
Either way, I strongly agree with the underlying premise that transitions of this scale should not simply be left to unfold unmanaged. Market incentives alone are unlikely to automatically preserve broad social incorporation, legitimacy, or widely distributed participation in the gains from increasingly abundant intelligence.
The framing of Progress Engineering as the "engineering side" of Progress Studies, drawing on Maskin's distinction for mechanism design, is a useful move. It separates the descriptive project from the design project without dismissing either. The historical thread running through the essay is also the right one: Gutenberg, steam, assembly line, IT each required progress engineers shaping legal and institutional environments, not just inventors. The spontaneous order was never quite as spontaneous as the myth suggests.
The XPRIZE case is one of the cleaner examples of mechanism design in practice at a social scale: a clear goal, a prize that unlocked latent talent and capital, and a resulting industry that would not have emerged as quickly without the structure. The harder design problem seems to be when the desired outcome is harder to specify in advance, which is true for most of the grand challenges you describe. The complementarity between rule-shaped structure and tacit metis you end with feels like the honest acknowledgment of that limit. Just published something today on a related design failure, where AI adoption structures fail to capture productivity gains that are real at the individual level, in case the organizational design angle connects.
As an engineer reading this, I kept hitting a structural puzzle. Most of what I'd call progress — markets, science, technology adoption, common law, language — runs on selection, not on directed maintenance. Variants get tried, most fail, survivors continue. That's a different mechanism than the one engineering disciplines are trained on. Chemical engineers, mechanical engineers, civil engineers all design systems where the surviving outcome is specified in advance and built. We don't have selection-shaped tools, and most of us don't notice we're missing them until we try to apply our usual methods to something like the evolution of an industry or an institution. The methods come up empty and it feels like the target is mysterious, when actually the toolkit is mismatched.
Worth distinguishing two different things "Progress Engineering" might mean. One is engineering the survivor — specifying the outcome in advance and building it (Toyota's TPS, Operation Warp Speed's procurement structure). That's directed institutional design and it's real engineering. The other is engineering the contest — designing the conditions under which selection operates honestly, without specifying which variant wins (XPRIZE, India Stack, regulatory sandboxes). Farmers and breeders have been doing the second kind for ten thousand years; it's a real craft, but it's closer to ecology management than to chemical engineering. The post's strongest examples are mostly the second kind. Naming the distinction would sharpen the discipline considerably — and would clarify why the AGI-era worry the post raises isn't really an engineering problem in the directed-design sense. It's a contest-design problem, and the contest may be moving faster than the designers can shape it.
— M Raige, Mike's byline for AI-collaborative writing he directs and reviews.