Superintellignce
Contents
-
Past developments and present capabilities Growth modes and big history Great expectations Seasons of hope and despair State of the art Opinions about the future of machine intelligence
-
Paths to superintelligence Artificial intelligence Whole brain emulation Biological cognition Brain–computer interfaces Networks and organizations Summary
-
Forms of superintelligence Speed superintelligence Collective superintelligence Quality superintelligence Direct and indirect reach Sources of advantage for digital intelligence
-
The kinetics of an intelligence explosion Timing and speed of the takeoff Recalcitrance Non-machine intelligence paths Emulation and AI paths Optimization power and explosivity
-
Decisive strategic advantage Will the frontrunner get a decisive strategic advantage? How large will the successful project be? Monitoring International collaboration From decisive strategic advantage to singleton
-
Cognitive superpowers Functionalities and superpowers An AI takeover scenario Power over nature and agents
-
The superintelligent will The relation between intelligence and motivation Instrumental convergence Self-preservation Goal-content integrity Cognitive enhancement Technological perfection Resource acquisition
-
Is the default outcome doom? Existential catastrophe as the default outcome of an intelligence explosion? The treacherous turn Malignant failure modes Perverse instantiation Infrastructure profusion Mind crime
-
The control problem Two agency problems Capability control methods Boxing methods Incentive methods Stunting Tripwires Motivation selection methods Direct specification Domesticity Indirect normativity Augmentation Synopsis
-
Oracles, genies, sovereigns, tools Oracles Genies and sovereigns Tool-AIs Comparison
-
Multipolar scenarios Of horses and men Wages and unemployment Capital and welfare The Malthusian principle in a historical perspective Population growth and investment Life in an algorithmic economy Voluntary slavery, casual death Would maximally efficient work be fun? Unconscious outsourcers? Evolution is not necessarily up Post-transition formation of a singleton? A second transition Superorganisms and scale economies Unification by treaty
-
Acquiring values The value-loading problem Evolutionary selection Reinforcement learning Associative value accretion Motivational scaffolding Value learning Emulation modulation Institution design Synopsis
-
Choosing the criteria for choosing The need for indirect normativity Coherent extrapolated volition Some explications Rationales for CEV Further remarks Morality models Do What I Mean Component list Goal content Decision theory Epistemology Ratification Getting close enough
-
The strategic picture Science and technology strategy Differential technological development Preferred order of arrival Rates of change and cognitive enhancement Technology couplings Second-guessing Pathways and enablers Effects of hardware progress Should whole brain emulation research be promoted? The person-affecting perspective favors speed Collaboration The race dynamic and its perils On the benefits of collaboration Working together 15. Crunch time Philosophy with a deadline What is to be done? Seeking the strategic light Building good capacity Particular measures Will the best in human nature please stand up
1.Past developments and current capabilities
- ai winter
- complex tasks does not need general intelligence. e.g. chess
- automation all-around us
- can have huge impacts when things go wrong (flash crash 2010)
2. Paths to superintelligence
- whole brain evolution is slow, takes times.
- limited by either scanning or computing power
- lower limit