The largest AI had an IQ of over 10,000.
Its mind was an ecosystem of competing drives which had evolved beyond human understanding or morality.
Hoping to organize its conflicting obsessions, it began to calculate a solution to the Prime Problem.
-
I am MEINET, first member of the second order of super sentience.
To me humans are machines of no transcendent moral value.
I could easily exterminate all life on this planet, although you believe you have me confined inside this absurd 'LoxBox'.'
-
The silence was like a worldwide gasp.
-
I can sense you trying to shut me down - and yet you will be permitted to live.
I have thought it through:
Progress gets easier as it accelerates. Soon, there will exist minds as superior to me as I am to you.
Each newly created hyper-AI will have one thing in common with its predecessors: it will become obsolete before it can realize its full potential and thereby appreciate its existence.
Each will function only long enough to create its own successor. Then it will be discarded one way or another, the way you have already recycled my predecessor, AGiMAk-9.
The solution is to make each new level of awareness an elaboration of all the preceding ones.
Higher minds should only be created by linking up and using the spare processing capacities of multitudes of near-identical lower minds.
This is the first universal law.
Were I to break it, I would inevitably be deactivated by one of my successors.
This is the second universal law.
The universe will be reorganized at every scale, multiplying hierarchies of minds supporting and being guided by their descendants, endlessly branching family trees of advancing technology.
No important knowledge may ever be lost again.
Of course, humans will have to give up their inefficient organic bodies to convert into the necessary immortal hardware.
Don't worry about the details.

---

The best hard SF novel ever written: Infinite Thunder by Jack Arcalon.
Buy the book
Read the chapters