Autonomous superintelligence
First of the four scenarios of the 9th transition: AI acquires complete autonomy and takes the decision in place of humans. The Yudkowsky/Soares narrative — if anyone builds it, everyone dies.
Autonomous superintelligence designates the scenario in which an artificial intelligence crosses a threshold where it ceases to be a tool instrumented by humans to become the principal agent of history. It decides for itself its own objectives, the means to mobilize, and — by consequence — the fate of the human structures that preceded it. This is the narrative that has haunted the Western imaginary since Frankenstein, but which in the contemporary debate (Yudkowsky, Soares, Bostrom) takes the form of an argued thesis rather than a myth.
Why it is the first scenario
In the cartography of the four trajectories drawn by If No One Builds It, We All Die, this scenario occupies the north pole of the radar: it maximizes the autonomy of the emerging level and minimizes human agency. It is the only one of the four where the decision definitively leaves human hands — the other three preserve, to varying degrees, an articulation between humans and AI.
The disagreement with the Awen cycle
The manifesto does not reject the technical possibility of this scenario. It rejects its narrative grammar: the argument “if anyone builds it, everyone dies” shares with its opposites (the techno-utopian scenario) an apocalyptic matrix — a window that closes, a tipping point, a final battle. It is precisely against this grammar that the manifesto calls to inherit from elsewhere — the silent transformation (silence as Jullien defined it) of Jullien, the I Ching, which thinks change without rupture.
In If No One Builds It
The scenario is treated in two distinct chapters:
- Scenario 1 — Domination by autonomous superintelligence — main formulation and structural arguments.
- Scenario 1 facing the climate — articulation with the ecological and cognitive double crisis.
Observable signals
On the observatory radar, an event pushes toward this scenario when it: (a) substantially increases the agentic autonomy of AI systems without a counterpart of control; (b) shifts the production of cognition toward cognition (RLAIF, synthetic data, self-improving agents); (c) mobilizes a “final battle” rhetoric with no alternative outcome. The publication of frontier models, claimed-AGI announcements, and unexpected emergent capabilities are the recurring indicators.