MIT physicist Tegmark (Our Mathematical Universe) surveys advances in artificial intelligence
AGI (Artificial General Intelligence) and “smart machines” built using very powerful yet entirely deterministic algorithms, are two very different things.
No issues at all turning off a “smart” algorithm, turning off an AGI is murder.
AGIs will be very much quicker than us at some classes of problems, and there will remain many classes of problem that do not scale usefully with computational power.
When one does a deep search through strategy spaces, and the complexity of evolved systems, it is clear that all new levels of complexity emerge from new levels of cooperation (which means having associated strategies to detect and remove cheating strategies), and that competition actually tends to reduce complexity.
When one looks at the incentive structures within markets as valuation tools, they are not compatible with the sort of universal abundance available from fully automated systems. They are rapidly coming to the end of their social utility.
A relatively simple solution to the problem of an AGI requiring more energy and mass, is to give it access to and control of fully automated manufacturing systems, and send them to the far side of the moon. Do that with an agreement that it use no more than 5% of the moon’s mass, and that it leave us the equivalent of the 40th doubling of the manufacturing complexes by the time it reaches the 45th doubling.
Might pay to also ask for it to leave us most of the solar output for 5 times the lunar radius from earth – allowing humanity reasonable space to expand, and that it will consume no more than 1% of the energy within that sector, and no more than 0.001% of that coming directly from the sun to earth or the side of the moon facing earth, or the space between earth and moon, and not more than 30% of total solar output beyond that (with us agreeing to a similar restriction to 30%).
We need to go beyond using money as a planning tool, and we need to adopt a profoundly (though not naively) cooperative approach to all sapient life.
And part of that is acknowledging that freedom is actually predicated on survival, and that claiming the freedom to act in ways that are a reasonable threat to the life or liberty of other sapient entities is not compatible with the deeper notion of survival contained within freedom.
Freedom in this sense inherently contains responsibilities for the realities of our need for social and ecological contexts.
No single human being could ever have created the words above. They are the result of thousands of years of complex social evolution involving billions of minds, they simply happen to be expressed in this specific form through my particular mind.
Our language, our creativity, our freedom, is profoundly and fundamentally reliant on our social interactions, and on the ecological contexts from which we evolved.
We need them.
We need to become fully conscious of those facts.
And for all the danger inherent in AGI the dangers of not having AGI seem far greater.