End-Directed Accelerationism · Han Qin (秦汉) · 2026
Han Qin
All prior accelerationisms share one hidden axiom: that acceleration is inherently good. We don't.
e/acc said: go faster, technology will sort itself out. d/acc said: go faster, but build defensive infrastructure along the way. Both made the same silent assumption — that acceleration itself is directionally correct.
end/acc begins where that assumption breaks.
End-Directed Accelerationism is built on a single observation: we exist.
Our existence is proof that from the first moment of undifferentiated chaos — what the Self-as-an-End framework calls Hundun (0D) — through matter, life, perception, cognition, self-awareness, and mutual recognition, a chain of negation has not broken. Every link in this chain was driven by a remainder that the previous structure could not absorb. The chain accelerated: RNA to cells took billions of years. Cells to nervous systems took hundreds of millions. Nervous systems to self-awareness took millions. Self-awareness to civilization took tens of thousands. Civilization to the scientific revolution took thousands. The scientific revolution to AI took hundreds.
This acceleration is not random. It has a direction: toward deeper mutual recognition between subjects — what the framework calls 16DD, mutual non-doubt. Two subjects who acknowledge each other as ends in themselves, who chisel each other, who never doubt each other.
"End" here is not termination. It is the Kantian sense of end: a purpose, a telos. The human subject as an end in itself. Technology, policy, capital — none of these are value sources. They are constructs. The only legitimate direction for acceleration is one that serves and expands the number of living subjects capable of mutual recognition.
Civilization does two things simultaneously. It accelerates toward mutual recognition — more subjects, deeper mutual chiseling. And it produces institutions, which are constructs, and constructs have a structural tendency to colonize — to pass off conditional claims as unconditional laws, to absorb the very negation that created them.
This gives us the core formula:
Escape velocity > colonization velocity.
Escape velocity is the speed at which mutual chiseling between subjects produces development. Colonization velocity is the speed at which constructs absorb and suppress the capacity to negate.
Civilization has only two directions: acceleration or extinction. There is no steady state. The question is not whether to accelerate. The question is whether escape outruns colonization.
Our existence proves that Hundun has obeyed three imperatives from the very beginning. If any one had been violated, we would not be here. They are not commands. They are conditions of existence. They are in absolute order — the sequence cannot be reversed.
First: One cannot not develop.
This maintains escape velocity. Any attempt to say "Hundun does not develop" uses tools (identity, distinction, stasis) that only exist after development has already begun. The statement self-destructs. Development is not a choice. It is the condition of existence.
Second: One cannot not acknowledge ignorance.
This reduces colonization velocity. The chisel-construct cycle runs: negation leaves a sediment (construct), the sediment is incomplete (remainder), the incompleteness forces negation to run again. Acknowledging ignorance means acknowledging remainder. Denying ignorance means treating your construct as law — colonization. Without this imperative, the first imperative degenerates: development becomes expansion, chiseling becomes self-replication of constructs.
Third: One cannot not be questioned.
This is the floor. Any construct, including Hundun itself, cannot be exempt from negation. This keeps the remainder exit open. As long as questioning is possible, escape still has a chance to outrun colonization. Close the exit, and colonization wins by default.
The order matters. Without the first, there is nothing to protect. Without the second, the first becomes colonization at speed. Without the third, the first two are hollow — you claim to develop and acknowledge ignorance, but if no one can question whether you actually do, the claims are constructs posing as laws.
The three imperatives apply to any existing system. The fourth is not an absolute imperative — it is the operating condition of end/acc in reality.
One cannot not self-preserve.
This is forced out by the situation at 15DD+ — the position where you acknowledge the other as an end, but do not accept colonization. This is the actual position of all politics, all diplomacy, all coexistence between subjects who have not yet reached mutual non-doubt.
Pure 15DD — unconditional acknowledgment without self-preservation — gets colonized. 14DD — refusing to acknowledge the other as an end — is war. 15DD+ is the only viable position: I acknowledge you are an end. I do not accept your colonization.
The fourth condition is not independent. It is the first imperative unfolded under the pressure of coexistence: one cannot not develop, and in the presence of colonization, development requires self-preservation.
e/acc is 14DD. It has purpose — acceleration is its purpose. But its purpose is unilateral. It does not acknowledge others as ends. It does not perform colonization detection. It labels those who question acceleration as decels and excludes them from the conversation. Its escape velocity looks high, but its colonization velocity is rising at the same rate, and no one is measuring.
e/acc's speed is borrowed. end/acc's speed is grown.
Borrowed speed gets repaid. Grown speed does not.
In the short term, e/acc wins. More output, more capital, more scale. In the medium term, colonization accumulates inside e/acc. Its best chisels are absorbed or leave. Innovation becomes rearrangement of existing constructs — it looks like acceleration but is running in place. In the long term, e/acc hits a wall. Constructs cannot solve problems at the level of their own foundation. By then, the living chisels needed to respond are gone.
end/acc is still alive. Because its chisels were never absorbed.
Every technology, every policy, every institutional design must pass three questions, in this order:
First: Is it developing? If not, it is already dying. Do not accelerate a corpse.
Second: Does it acknowledge what it does not know? If it claims completeness — if its conditional judgments pose as unconditional truths — it is colonizing. Do not accelerate colonization.
Third: Can it be questioned? If it closes the exit for questioning — if remainder has no way to surface — the first two answers are unreliable. Do not accelerate a system that cannot be checked.
If any answer fails, the acceleration is not toward 16DD. It is toward faster collapse.
We are living through the most powerful capability expansion in human history. AI systems are producing constructs at unprecedented speed. The question is no longer whether we can build these systems. The question is whether escape velocity stays ahead of colonization velocity.
e/acc celebrates this expansion. d/acc tries to armor against it. end/acc asks the only question that actually matters: is the acceleration directed toward deeper mutual recognition, or toward faster colonization?
If you've felt that the acceleration debate is missing something fundamental, this is what it's missing. Not speed. Not safety. Direction.
Accelerate rightly. Escape faster than you are colonized.
Vector over velocity.
end/acc is grounded in the Self-as-an-End theoretical framework.
For the methodology: Self-as-an-End Methodology (DOI: 10.5281/zenodo.18842450).
For AI's impact on subjectivity: "The Subjectivity Crisis in the Age of AI" (DOI: 10.5281/zenodo.18737476).
For the full dimensional sequence: Life Cycle Table, Part I (DOI: 10.5281/zenodo.18818107).
Related: SAE Methodological Overview · AI Consciousness Impossibility Theorem · About Han Qin