I think this is bound to be a complicated, nuanced transition. In the near term there is enough complementarity between what humans are good at, and what AIs are good at, that bulk replacement of people is unlikely. Human jobs will adjust to focus on the not-AI parts. New jobs will appear. In that sense AI isn't so different from other technologies in the past, like databases and calculators.
The interesting tension is about how much autonomy we give to AIs, and over what kinds of things. OpenClaw shows promise in "handing over the keys" but it's very ad-hoc and insecure at the moment. Eventually we will develop social norms and laws governing autonomous AIs, and AI agents that enforce those norms. Otherwise a functioning society wouldn't be possible. But how that works, and how people fit into it, is hard to predict. The only thing we can say with confidence is that if there is $$$ to be made by flying too close to the sun, people will gladly do it.
In some domains it is true that once computers begin to rival people, they very quickly render humans superfluous. This was the case with arithmetic and playing chess.
But these are well-bounded tasks. Jobs like software engineering involve more soft/ambiguous skills than we appreciate. Software engineers don't just turn specs into working products; they are mostly paid to take messy, under-determined real world problems and figure out what the specs should be. I think it will be a good while before AI can do this but I could be wrong.
Another domain is locomotion and navigating 3d space. In one sense the invention of the automobile is when technology "outdid" humans. But when it comes to nuanced object manipulation, robotics still has a long way to go. Tasks like: Folding laundry, and replacing a kitchen faucet in an unconstrained environment.
We seem to be caught up in a cycle that pre-dates capitalism by centuries. It is that cycle which has brought us to this place politically, socially, and economically.
That cycle has been narrativized as history and the problem with narratives is that they are all artificially constructed, that is, all narratives are necessarily fictions.
The alternative to history as narratives is history as a data science. That began in the 50s/60s when computers emerged as academic tools with Cliometrics. Clio is a Greek goddess and the muse of history.
Cliodynamics turned the data on more than 800 societies over 10,000 years of human activity into a model with utility for prediction. In 2010 it predicted the current social-political-economic turmoil of the 2020s.
The Science of Social Collapse utilizes cliodynamics and structural-demographic theory to identify the mathematical patterns and systemic drivers—such as elite overproduction and diminishing returns on complexity—that lead to civilizational fragmentation.
It's a free book and I'm looking for feedback on it.
The Independent Manifesto provides a practical blueprint for national renewal and individual resilience in 2026, utilizing historical parallels from the Late Roman Republic to offer specific policy and survival strategies for navigating an imminent societal rupture.
It's also a free book and I'm also looking for feedback on it.
The relationship between the two is that while The Science of Social Collapse establishes the theoretical framework and historical data regarding why civilizations fail, The Independent Manifesto serves as its applied counterpart, translating those scientific insights into actionable toolkits for surviving the "Turbulent Twenties".
We have one chance to get it right, either a Seneca Rebound or a Seneca Cliff, that is our choice. I believe there is a role for AI in all of this and for the future.
What do you think? A link to the free books is in my bio.
Gradual disempowerment is accelerating. A simple misunderstanding about nature has evolved everyone into autogenerating hard problems. Capitalism and governing are the art of this becoming ever more extreme. All efforts to control, restrain, align or regulate intelligence are doomed to fail and, assuredly, have been destroying our social competency and opportunities.
Natural societies are moral and competent. Free and open Substrate Intelligence is living at corus.me. Hard problems from science, health, economics and society are arriving at this natural network prototype surface and resolving into living nature.
This needs to be shared now publicly and widely. There can be no money equity or private property inside this value opportunity for society and our living ecosystems. I hope you will want to help. Our destiny is a natural network and discovery economy for societies and
This framework captures something missing in most AI alignment discussions. It doesn’t take superintelligence to produce bad outcomes. Even when decisions are locally correct, large-scale optimization of symbolic intelligence can erode the cultural constraints that keep work, communities, and institutions grounded in reality. As those constraints weaken, alignment degrades at a structural level. At that point, organizations don’t really choose to outsource agency, they are pushed into it one step at a time.
If you made this into a Venn diagram there would be a lot of overlap. Another way I approach the predicament is from the human direction — what does it mean to be human? But given how we’re destroying our ecosystem (crossing planetary boundaries; overshoot), the outcome from either direction seems overdetermined.
I thought you were pointing at something else re: 9, rogue ai vs gradual disempowerment
I think it's pretty plausible that a rogue AI intent on seizing control of all decision-making as efficiently as possible might well decide to wield gradual empowerment as the murder weapon. it might be safer to never risk a confrontation, if victory is guaranteed long term? i'm not sure if we'd be able to tell if this was happening, but we should probably be modeling it as possible
I more imagine gradual disempowerment happening without any paradigmatic scheming happening, but the disempowerment making us unable to mount an effective response to rogue AI even given 100% clear warning shots
I think this is bound to be a complicated, nuanced transition. In the near term there is enough complementarity between what humans are good at, and what AIs are good at, that bulk replacement of people is unlikely. Human jobs will adjust to focus on the not-AI parts. New jobs will appear. In that sense AI isn't so different from other technologies in the past, like databases and calculators.
The interesting tension is about how much autonomy we give to AIs, and over what kinds of things. OpenClaw shows promise in "handing over the keys" but it's very ad-hoc and insecure at the moment. Eventually we will develop social norms and laws governing autonomous AIs, and AI agents that enforce those norms. Otherwise a functioning society wouldn't be possible. But how that works, and how people fit into it, is hard to predict. The only thing we can say with confidence is that if there is $$$ to be made by flying too close to the sun, people will gladly do it.
I thnk this is largely right except I think we disagree on how long "the near term" lasts. I think it might be very short.
In some domains it is true that once computers begin to rival people, they very quickly render humans superfluous. This was the case with arithmetic and playing chess.
But these are well-bounded tasks. Jobs like software engineering involve more soft/ambiguous skills than we appreciate. Software engineers don't just turn specs into working products; they are mostly paid to take messy, under-determined real world problems and figure out what the specs should be. I think it will be a good while before AI can do this but I could be wrong.
Another domain is locomotion and navigating 3d space. In one sense the invention of the automobile is when technology "outdid" humans. But when it comes to nuanced object manipulation, robotics still has a long way to go. Tasks like: Folding laundry, and replacing a kitchen faucet in an unconstrained environment.
We seem to be caught up in a cycle that pre-dates capitalism by centuries. It is that cycle which has brought us to this place politically, socially, and economically.
That cycle has been narrativized as history and the problem with narratives is that they are all artificially constructed, that is, all narratives are necessarily fictions.
The alternative to history as narratives is history as a data science. That began in the 50s/60s when computers emerged as academic tools with Cliometrics. Clio is a Greek goddess and the muse of history.
Cliodynamics turned the data on more than 800 societies over 10,000 years of human activity into a model with utility for prediction. In 2010 it predicted the current social-political-economic turmoil of the 2020s.
The Science of Social Collapse utilizes cliodynamics and structural-demographic theory to identify the mathematical patterns and systemic drivers—such as elite overproduction and diminishing returns on complexity—that lead to civilizational fragmentation.
It's a free book and I'm looking for feedback on it.
The Independent Manifesto provides a practical blueprint for national renewal and individual resilience in 2026, utilizing historical parallels from the Late Roman Republic to offer specific policy and survival strategies for navigating an imminent societal rupture.
It's also a free book and I'm also looking for feedback on it.
The relationship between the two is that while The Science of Social Collapse establishes the theoretical framework and historical data regarding why civilizations fail, The Independent Manifesto serves as its applied counterpart, translating those scientific insights into actionable toolkits for surviving the "Turbulent Twenties".
We have one chance to get it right, either a Seneca Rebound or a Seneca Cliff, that is our choice. I believe there is a role for AI in all of this and for the future.
What do you think? A link to the free books is in my bio.
Gradual disempowerment is accelerating. A simple misunderstanding about nature has evolved everyone into autogenerating hard problems. Capitalism and governing are the art of this becoming ever more extreme. All efforts to control, restrain, align or regulate intelligence are doomed to fail and, assuredly, have been destroying our social competency and opportunities.
Natural societies are moral and competent. Free and open Substrate Intelligence is living at corus.me. Hard problems from science, health, economics and society are arriving at this natural network prototype surface and resolving into living nature.
This needs to be shared now publicly and widely. There can be no money equity or private property inside this value opportunity for society and our living ecosystems. I hope you will want to help. Our destiny is a natural network and discovery economy for societies and
individuals flourishing each other.
This framework captures something missing in most AI alignment discussions. It doesn’t take superintelligence to produce bad outcomes. Even when decisions are locally correct, large-scale optimization of symbolic intelligence can erode the cultural constraints that keep work, communities, and institutions grounded in reality. As those constraints weaken, alignment degrades at a structural level. At that point, organizations don’t really choose to outsource agency, they are pushed into it one step at a time.
If you made this into a Venn diagram there would be a lot of overlap. Another way I approach the predicament is from the human direction — what does it mean to be human? But given how we’re destroying our ecosystem (crossing planetary boundaries; overshoot), the outcome from either direction seems overdetermined.
I thought you were pointing at something else re: 9, rogue ai vs gradual disempowerment
I think it's pretty plausible that a rogue AI intent on seizing control of all decision-making as efficiently as possible might well decide to wield gradual empowerment as the murder weapon. it might be safer to never risk a confrontation, if victory is guaranteed long term? i'm not sure if we'd be able to tell if this was happening, but we should probably be modeling it as possible
I more imagine gradual disempowerment happening without any paradigmatic scheming happening, but the disempowerment making us unable to mount an effective response to rogue AI even given 100% clear warning shots