[ExI] Digital identity

[Anders Sandberg]

On 26/04/2013 13:54, BillK wrote:
> In the West, to a great extent the rich do control governments.
> Freedom means anyone can be bought and those that can't be bought get
> moved.

I assume you are merely donning some fashionable cynicism here, rather 
than actually talking political science or sociology?

There is an interesting question in how influence scales with wealth. 
Clearly richer people do have bigger social impact, but much of it is 
contact networks rather than through wealth itself. As a thought 
experiment, how many ordinary Americans (net worth per family around 
$200,000) would it take to balance the wishes of Bill Gates (67 
billion)? If the influence as function of wealth is linear, you should 
expect that 335,000 opposing families (around 871,000 people) to be 
required. If it is sublinear, then smaller interest groups can 
successfully thwart Bill. An interesting exercise is to check the sizes 
of the constituencies that have successfully thwarted various interests; 
obviously there is much more to each case, but it might make a fun 
project in political economy.

It is also worth noting that the richest individuals today are typically 
far poorer than the governments where they live, which was not always 
true: the Medicis had more money than the Republic of Florence, Jakob 
Fugger II bankrolled the Holy Roman Emperor and the Pope. But today, 
Mexico has a budget on the order of $290 billion, and Carlos Slim has 73 
billion. Bill gates merely got 67 billion compared to a US budget of 3 
trillion. If you believe influence scales linearly or superlinearly with 
available money, the governments will clearly be able to control the 
rich guys.

> Forget growth rates. That's past history. Once you are a nano-scale 
> self-supporting virtual universe drifting in space, growth means 
> little to you. Communication also means little as the outside universe 
> has frozen due to your internal clock speedup. You might want to 
> communicate with other nearby uploads, but who knows???

The kind of people who want to live forever in the garden of Eden are 
typically not the kind of people who drive the economy and make history. 
Sure, there will be a bunch of little Edens floating around, but they 
will not matter much since the bulk of growth (economic, scientific, 
artistic) will be happening in the interconnected economy.

>
>> This depends on how scanning cost and computing cost scales. If scanning is
>> expensive compared to computing, then there will be few different minds but
>> they could run many of copies. So after the first crazy test subjects you
>> get rich uploads, who presumably do not want to copy too much. But sooner or
>> later you will get a widely copied worker upload, and then things get very
>> different. If computing is expensive, then there will be few copies and the
>> wealthy will dominate... as long as computing remains expensive. Which may
>> not be long: given Moore's law timescales, if you can afford one upload now,
>> in ten years you can run thousands.
>>
> The problem is that uploads process much faster than humans. Once
> uploaded, to them the rest of the world appears to stop. This is the
> same problem as a runaway AGI. Once it happens, they choose what
> happens next.

The speed depends on available computer power, and this depends on the 
particular scenario. If the emulation technology arrives first, but 
computers are not yet cheap/fast enough uploads will be few and slow, 
gradually picking up speed. If you have hardware overhang due to late 
emulation technology, you get fast/many uploads quickly.

If the amount of computing power grows as exp(kt), and you have upload 
populations growing as exp(lt) (where k and l are constants), the 
average speed will grow as exp[(k-l)t]: Even a very rapid Moore's law 
can be eaten up by a rapidly growing upload population. The exact growth 
depends on the economic benefit of adding an extra upload (which will 
have a cost ~exp(-kt)). If we assume a constant benefit B, then the gain 
from each upload will be B-exp(-kt) - as soon as exp(-kt) goes below B 
you will start adding uploads as fast as you can, and the population 
will grow roughly as exp(kt), not gaining enormous speed except where it 
is economically very important. More realistically, if the benefit 
declines with the number of existing uploads, then the number of uploads 
will grow more slowly and they will be getting faster. A really proper 
economic model would then model the feedback between the growth of the 
economy, uploads, and the demand for more computers. If I remember the 
work of Robin and Carl right, it seems that population growth will tend 
to be the primary mode, but I should check this.




-- 
Anders Sandberg,
Future of Humanity Institute
Oxford Martin School
Faculty of Philosophy
Oxford University