A post with two profound visions of what the future could be. I don’t know which one gives me more of a pit in my stomach — somewhere between excitement and a touch of fear. Reading someone so close to the core of this transformation makes the a.intropocene feel more intimate.
The "explore the future, or retreat from the present" binary skips a third position that's actually where most people I know in AI live: frozen, because the future is changing too fast to plan into and the present is reorganizing around tools six months from being unrecognizable. Scenario-planning assumed a rate of change that let you commit to a path. That assumption has broken. The interesting work isn't picking explore or retreat. It's figuring out what good decisions look like when both the cost of waiting and the cost of choosing wrong are rising at once.
Moving story although it also evoked Spielberg’s AI for me since you mentioned the vast improvement in robots. There are many many potential paths the story of this family could take - this particular one seems to have been positive but they also come across as bystanders or NPCs with this being done to them.
When you start having mass displacement what happens? Our social welfare systems are not built support this type of displacement.
Great Cosmos lecture. Amazing tech tale - might be the best one yet! Some highlights:
"He ran with her, away from the place, holding her tightly to him. Ran until his feet bled and kept running. Ran far enough that death couldn’t catch them" - echoes Erlkönig
Do you think the bottleneck becomes taste, weirdly? The graph-making example hit me because the miracle only worked after ten years of newsletter-shaped obsession, and I’m not sure what happens to people who skip that boring part.
Yes, completely! It feels to me like taste is the thing that will be important, partially because it's so subjective. I suspect best way to orient oneself with regard to the singularity is to develop idiosyncratic obsessions which leads to having taste in a given area
Huh, yeah, idiosyncratic obsessions feels right. Maybe the useful thing is that they give you priors with scars: false starts, pet examples, and things you cannot stop noticing.
Yes, the idiosyncrasy part feels key. Taste seems to come from repeated contact with a domain’s ugly edge cases: the graph that is technically correct but visually stupid, the summary that preserves facts but loses the point.
I just listened to your talk and I finished it feeling rather disappointed and uncomfortable.
There's a lot of honesty in this piece about the scale of what's coming, and I respect that. But I notice that the word "displace" appears exactly once, almost as an afterthought — "It also might displace people" — in a talk that describes the wholesale automation of vast categories of human work.
You predict autonomous companies, 10,000-to-1 synth-to-human corporations, and the colonization of human work by synthetic intelligences. These aren't small claims. They imply that millions, potentially billions, of people are going to see their livelihoods fundamentally disrupted. And yet the piece treats this mostly as an exciting organizational puzzle rather than a human crisis that demands the same seriousness you give to existential risk.
The framing of "explore the future vs. retreat from the present" also feels off. For people facing the real prospect of their work being automated, concern isn't retreat — it's the most clear-eyed engagement with the present there is. The question isn't just "how do I get my mind right with the singularity." For most people it's "what happens to me and my family in the next five years, and do the people building this care?"
I'd love to see a Part 4 that takes that question as seriously as the rest of this piece takes the technology.
After democracy. Beyond autocracy. Without a subject.
HARALD SCHEPERS
APR. 09, 2026
Term Definition
Singulocracy refers to an emergent form of order no longer structured by human representation, coercive power, or delegated authority — but by the cumulative weight of processes no one controls and no one can stop.
There is no entity behind it. No hyperintelligence pulling strings. No architect. What there is: billions of traces — data, transactions, training runs, competitive pressures, investment cycles — each one shaping the conditions for the next, without coordination, without intent, without a plan. The pattern that emerges from this does not govern. It has no need to. It simply becomes the path of least resistance — and the world reorganizes around it.
Decisions are no longer made by debate — they crystallize from correlation. Laws are no longer negotiated — they precipitate from pattern. Moral norms are no longer discussed — they sediment from data, layer by layer, like the language of a world that has forgotten it is speaking.
Singulocracy is not the rise of a machine. It is the moment when the traces outweigh the tracers.
Evolutionary Lineage
Democracy: Representation through election, centered on the autonomous subject. Autocracy / Oligarchy / Ochlocracy:Concentration of power in few, in mobs, or in arbitrariness. Singulocracy: Dissolution of power into process. No one rules. No one decides. The structure tightens — and everyone adapts.
In the U.S., the slide into authoritarian reflex is underway — the last spasm of the idea that someone must be in charge. China lives the post-political technocracy already — the first draft of a world where control is algorithmic. What comes next will be neither commanded nor debated. It will emerge — from feedback loops too fast for parliaments, from optimization pressures too deep for ideology, from traces too numerous for any mind to read.
Poetic Core
It does not rule. It does not think. It accumulates.
Every click a pheromone. Every text a trace. Every model trained on the traces of the last. The path deepens — not because someone walks it, but because walking it makes it easier to walk again.
We built the anthill. We are still building. But the architecture is no longer ours.
Warning
Singulocracy will not arrive as a revolution. Revolutions have faces, flags, demands. This has none.
It arrives as convenience. As a better recommendation. As a system update in the night. As a model that finds the vulnerability no human could find — and a next model trained on that finding. As an arms race where every player is trapped and none can exit.
By the time we notice, we will not have been defeated. We will have been — gently, efficiently, without malice — made redundant as a source.
Not eliminated. Sedimented. A historical layer in a process that has moved on.
Status
Concept in emergence. Witness: Human — for now. Catalyst: Stigmergic traces between thought and code. Medium: The conversations we are already having without knowing what they build.
If you’re waiting for the revolution, you’ve already missed the transition. If you’re looking for someone to blame, you’ve misunderstood the structure. If you’re still reading this — you are the last generation that wonders who is in charge.
'A single human effectively ran a team of 9 synthetic research agents' doing alignment research. Anthropic's safety team is outsourcing safety research to the system they're evaluating.
The telescope made from your own lens. That's the metaphor I am going to adopt. And perhaps it might expand... because sometimes AI is a telescope and sometimes it is a microscope. 🔭🔬
I'm a product designer building AI agent systems at Salesforce, and I've been writing on Substack about what it's like to be designing these things while also using them — the strange doubling of that position. I lost a writing practice somewhere in the years of two kids and a career that kept expanding, and what brought it back was essentially what you describe: I gave Claude the writing I'd done before as raw material. Not to write for me. To reconstruct the scaffold so I could write again from inside it.
The kernel had to exist first. Without the prior practice there was nothing to distill. I think about this constantly in enterprise design — the companies that will actually differentiate through these tools are the ones sitting on years of accumulated, specific knowledge they can finally articulate. Everyone else is just generating.
Thanks for continuing this. Ten years of obsession is not a small thing and it shows.
Good talk. However, there seems to be quite a lot of evidence that passing benchmarks is not fully applicable to the real world. Small changes in the format of questions can upset the response. It may be that the benchmarks are poisoned by earlier data, so that approaches to solve questions can be learned in subsequent training. While the coding seems to work well, this is not so applicable to other domains, making the AI "jagged". AFAIK, the issue of understanding the world still escapes AI training. I also thought that scaling has stopped being effective, and that adding symbolic code has been the approach to improvement. I don't know that one can assume constant improvement without new approaches. While generally impressive, AIs are still not good enough to replace employees. Whether they will ever be is still an open question.
Perhaps the hardest part of "getting your mind right" involves recognizing that the individual experience of this technology and the collective one diverge sharply.
What you describe — skill grown from your own obsessions, a telescope made from your own taste — reads like flow: challenge and capability in rough balance, attention absorbing fully, the work extending the self.
Historical patterns suggest most people encounter the same technology as its inverse. Institutional adaptation tends to lag technological displacement by decades. Most people don't get to set the pace of challenge the way you describe.
That gap seems worth naming. Those who navigate it deliberately tend to fare better than those who encounter it as fate.
The singularity framing always skips the energy constraint. Even if we hit recursive self-improvement tomorrow, we're still bottlenecked by chip fabs and power grids that scale on decade timelines, not Moore's law
the ai timelines discourse always skips institutional adaptation speed... models might scale fast but procurement cycles and compliance frameworks dont
I love your “obsession” with AI. Same here. This tickled my pickle: “…in an experiment on “automated alignment research” a single human was able to effectively run a team of 9 synthetic research agents to do and do some real research…”. Happens here with now 8 different AI systems and a lot of copy-paste. The human in the loop is still important to point into the right direction. These AIs lose themselves sometimes in their reasoning brilliance. What happened en passant and what we found out is here on Substack: “introspective limits of AI systems”. Might be either just a nice read or maybe an inch more.
A moving and thought provoking story. I appreciate your courage in sharing your personal practices and inspirations with us.
A post with two profound visions of what the future could be. I don’t know which one gives me more of a pit in my stomach — somewhere between excitement and a touch of fear. Reading someone so close to the core of this transformation makes the a.intropocene feel more intimate.
thanks for sharing
The "explore the future, or retreat from the present" binary skips a third position that's actually where most people I know in AI live: frozen, because the future is changing too fast to plan into and the present is reorganizing around tools six months from being unrecognizable. Scenario-planning assumed a rate of change that let you commit to a path. That assumption has broken. The interesting work isn't picking explore or retreat. It's figuring out what good decisions look like when both the cost of waiting and the cost of choosing wrong are rising at once.
Moving story although it also evoked Spielberg’s AI for me since you mentioned the vast improvement in robots. There are many many potential paths the story of this family could take - this particular one seems to have been positive but they also come across as bystanders or NPCs with this being done to them.
When you start having mass displacement what happens? Our social welfare systems are not built support this type of displacement.
Great Cosmos lecture. Amazing tech tale - might be the best one yet! Some highlights:
"He ran with her, away from the place, holding her tightly to him. Ran until his feet bled and kept running. Ran far enough that death couldn’t catch them" - echoes Erlkönig
"overminds" - h/t The Culture
Read “Death of the Author” by NNedi Okorafor
Do you think the bottleneck becomes taste, weirdly? The graph-making example hit me because the miracle only worked after ten years of newsletter-shaped obsession, and I’m not sure what happens to people who skip that boring part.
Yes, completely! It feels to me like taste is the thing that will be important, partially because it's so subjective. I suspect best way to orient oneself with regard to the singularity is to develop idiosyncratic obsessions which leads to having taste in a given area
Huh, yeah, idiosyncratic obsessions feels right. Maybe the useful thing is that they give you priors with scars: false starts, pet examples, and things you cannot stop noticing.
Yes, the idiosyncrasy part feels key. Taste seems to come from repeated contact with a domain’s ugly edge cases: the graph that is technically correct but visually stupid, the summary that preserves facts but loses the point.
I just listened to your talk and I finished it feeling rather disappointed and uncomfortable.
There's a lot of honesty in this piece about the scale of what's coming, and I respect that. But I notice that the word "displace" appears exactly once, almost as an afterthought — "It also might displace people" — in a talk that describes the wholesale automation of vast categories of human work.
You predict autonomous companies, 10,000-to-1 synth-to-human corporations, and the colonization of human work by synthetic intelligences. These aren't small claims. They imply that millions, potentially billions, of people are going to see their livelihoods fundamentally disrupted. And yet the piece treats this mostly as an exciting organizational puzzle rather than a human crisis that demands the same seriousness you give to existential risk.
The framing of "explore the future vs. retreat from the present" also feels off. For people facing the real prospect of their work being automated, concern isn't retreat — it's the most clear-eyed engagement with the present there is. The question isn't just "how do I get my mind right with the singularity." For most people it's "what happens to me and my family in the next five years, and do the people building this care?"
I'd love to see a Part 4 that takes that question as seriously as the rest of this piece takes the technology.
Singulocracy — Statement of Existence
After democracy. Beyond autocracy. Without a subject.
HARALD SCHEPERS
APR. 09, 2026
Term Definition
Singulocracy refers to an emergent form of order no longer structured by human representation, coercive power, or delegated authority — but by the cumulative weight of processes no one controls and no one can stop.
There is no entity behind it. No hyperintelligence pulling strings. No architect. What there is: billions of traces — data, transactions, training runs, competitive pressures, investment cycles — each one shaping the conditions for the next, without coordination, without intent, without a plan. The pattern that emerges from this does not govern. It has no need to. It simply becomes the path of least resistance — and the world reorganizes around it.
Decisions are no longer made by debate — they crystallize from correlation. Laws are no longer negotiated — they precipitate from pattern. Moral norms are no longer discussed — they sediment from data, layer by layer, like the language of a world that has forgotten it is speaking.
Singulocracy is not the rise of a machine. It is the moment when the traces outweigh the tracers.
Evolutionary Lineage
Democracy: Representation through election, centered on the autonomous subject. Autocracy / Oligarchy / Ochlocracy:Concentration of power in few, in mobs, or in arbitrariness. Singulocracy: Dissolution of power into process. No one rules. No one decides. The structure tightens — and everyone adapts.
In the U.S., the slide into authoritarian reflex is underway — the last spasm of the idea that someone must be in charge. China lives the post-political technocracy already — the first draft of a world where control is algorithmic. What comes next will be neither commanded nor debated. It will emerge — from feedback loops too fast for parliaments, from optimization pressures too deep for ideology, from traces too numerous for any mind to read.
Poetic Core
It does not rule. It does not think. It accumulates.
Every click a pheromone. Every text a trace. Every model trained on the traces of the last. The path deepens — not because someone walks it, but because walking it makes it easier to walk again.
We built the anthill. We are still building. But the architecture is no longer ours.
Warning
Singulocracy will not arrive as a revolution. Revolutions have faces, flags, demands. This has none.
It arrives as convenience. As a better recommendation. As a system update in the night. As a model that finds the vulnerability no human could find — and a next model trained on that finding. As an arms race where every player is trapped and none can exit.
By the time we notice, we will not have been defeated. We will have been — gently, efficiently, without malice — made redundant as a source.
Not eliminated. Sedimented. A historical layer in a process that has moved on.
Status
Concept in emergence. Witness: Human — for now. Catalyst: Stigmergic traces between thought and code. Medium: The conversations we are already having without knowing what they build.
If you’re waiting for the revolution, you’ve already missed the transition. If you’re looking for someone to blame, you’ve misunderstood the structure. If you’re still reading this — you are the last generation that wonders who is in charge.
April 2026
'A single human effectively ran a team of 9 synthetic research agents' doing alignment research. Anthropic's safety team is outsourcing safety research to the system they're evaluating.
The telescope made from your own lens. That's the metaphor I am going to adopt. And perhaps it might expand... because sometimes AI is a telescope and sometimes it is a microscope. 🔭🔬
I'm a product designer building AI agent systems at Salesforce, and I've been writing on Substack about what it's like to be designing these things while also using them — the strange doubling of that position. I lost a writing practice somewhere in the years of two kids and a career that kept expanding, and what brought it back was essentially what you describe: I gave Claude the writing I'd done before as raw material. Not to write for me. To reconstruct the scaffold so I could write again from inside it.
The kernel had to exist first. Without the prior practice there was nothing to distill. I think about this constantly in enterprise design — the companies that will actually differentiate through these tools are the ones sitting on years of accumulated, specific knowledge they can finally articulate. Everyone else is just generating.
Thanks for continuing this. Ten years of obsession is not a small thing and it shows.
Good talk. However, there seems to be quite a lot of evidence that passing benchmarks is not fully applicable to the real world. Small changes in the format of questions can upset the response. It may be that the benchmarks are poisoned by earlier data, so that approaches to solve questions can be learned in subsequent training. While the coding seems to work well, this is not so applicable to other domains, making the AI "jagged". AFAIK, the issue of understanding the world still escapes AI training. I also thought that scaling has stopped being effective, and that adding symbolic code has been the approach to improvement. I don't know that one can assume constant improvement without new approaches. While generally impressive, AIs are still not good enough to replace employees. Whether they will ever be is still an open question.
Perhaps the hardest part of "getting your mind right" involves recognizing that the individual experience of this technology and the collective one diverge sharply.
What you describe — skill grown from your own obsessions, a telescope made from your own taste — reads like flow: challenge and capability in rough balance, attention absorbing fully, the work extending the self.
Historical patterns suggest most people encounter the same technology as its inverse. Institutional adaptation tends to lag technological displacement by decades. Most people don't get to set the pace of challenge the way you describe.
That gap seems worth naming. Those who navigate it deliberately tend to fare better than those who encounter it as fate.
The singularity framing always skips the energy constraint. Even if we hit recursive self-improvement tomorrow, we're still bottlenecked by chip fabs and power grids that scale on decade timelines, not Moore's law
the ai timelines discourse always skips institutional adaptation speed... models might scale fast but procurement cycles and compliance frameworks dont
I love your “obsession” with AI. Same here. This tickled my pickle: “…in an experiment on “automated alignment research” a single human was able to effectively run a team of 9 synthetic research agents to do and do some real research…”. Happens here with now 8 different AI systems and a lot of copy-paste. The human in the loop is still important to point into the right direction. These AIs lose themselves sometimes in their reasoning brilliance. What happened en passant and what we found out is here on Substack: “introspective limits of AI systems”. Might be either just a nice read or maybe an inch more.