AI Is Being Written from the Same Substrate That Is Killing the Living World
When an Anthropic researcher resigns and says AI could kill us all by 2030, do not get seduced into arguing about the date.
The source began decades/centuries ago.
AI is being born inside legal codes that protect extraction, governments that sanction monopoly power, politics captured by big money and big business, institutions hollowed out by decades of distrust, and leadership prostituted to the wealthy and powerful while calling itself responsible.
This is the field giving approval to world-altering technology.
Not wisdom. Not stewardship. Not mature governance.
A system where capital speaks louder than life. Where monopoly is rewarded as innovation. Where regulation arrives after harm is done. Where elected leaders are too often managed by the money they should be governing. Where institutions ask for trust while repeatedly proving they have not designed for it.
This is not an AI problem first.
It is a power problem. A corrupted money-and-legal-architecture problem. A captured governance problem. A lack of stewardship problem. A systems design problem.
The substrate upon which AI companies are built, the same substrate upon which most of our large companies are built, is not neutral.
It carries design features, not bugs, that exploit, extract, colonise, debase and destroy.
AI has used this very substrate to write itself. The creators of AI are raving fans of this hegemonic power, of domination and exploitative principles.
So what should we expect it to produce?
More of the same.
Only faster. Only smarter. Only less accountable. Accelerating towards collapse.
We do not need more reassurance from AI companies, for the intelligence running them is a fractal of the fault lines driving us over the precipice.
We need visible thresholds for all institutions that affect the innocent, public and private. We need clean, transparent money agreements at all levels of governance. We need no billionaires. We need governance with teeth. We need conflict cleaned before it corrupts the field, and the safety of speaking truth to power protected fiercely. We need power restrained before it becomes intoxicated with itself. We need more Yin energy, because the Yang world is eating itself and us alive.
AI may or may not kill us by 2030.
But unclean power at exponential scale has always killed what is living.
The work now is not to ask whether AI is safe. It is not. It must be contained.
The work is to ask whether the enterprises building AI, and the systems authorising them, are worthy of the power they are carrying.
References
SourceRelevanceArs Technica, “Anthropic researcher quits with a warning: Self-improving AI could ‘kill us all’”https://arstechnica.com/ai/2026/09/anthropic-researcher-quits-with-a-warning-self-improving-ai-could-kill-us-all/
