[ExI] Forking

[Anders Sandberg]

Keith Henson wrote:
> On Mon, Dec 26, 2011 at 5:00 AM,  Anders Sandberg <anders at aleph.se> wrote:
>   
>> Forking is not necessarily driven by a drive to reproduce. It might be
>> economic, it might be intellectual, it might even be religious - as soon
>> as there is somebody with pro-forking views a lot of forking is likely
>> unless the cost of computing space is very high.
>>     
>
> To me this discussion has the flavor of banging your head on a wall or
> putting beans up your nose.
>   

Ah. I think I see the fundamental problem. We likely have different 
views of personal identity, and that colors our perceptions. I regard 
myself as the equivalence class of all sufficiently Anders-like 
processes, and this (together with my value theory) make me think 
forking can add value. Your identity and value theories likely do not 
see it as useful as I do.


> Resources will always be limited.  Forking will on average cost a
> person (or society) half the computation space they had.  So after a
> fork, you run half as fast or are half as smart and half as rich.
> Someone with a madness for forking will soon have no more hardware per
> second than an Apple II.  There may be a lot of you, but at the rate
> they think it hardly matters.
>   

If I fork myself, there will now be twice as much me-experience, twice 
the amount of human capital and twice as many entities sharing my goals. 
This has a certain value to me. It is also valuable to others if I have 
useful skills - now these services can be bought a bit more cheaply, and 
at a greater quantity or speed. This later value equates to income: two 
forks of me can under these circumstances earn more than one of me. 
Also, even if the forks earn less per unit of time they can still save 
earnings and invest them.

There are interesting issues in fork economics, for example whether it 
is better to be twice as fast as twice as many, and how much the cost of 
computing substrate will be relative to the median mind salary. I think 
there are many domains where the advantages go in different directions.


> I can't think of any economic or intellectual reason that would not be
> served equally well by spinning off a temporary thread that would
> merge back in to the main stem.  As for the Church of the Fork . . . .
>
> The inverse of forking is merging.  That could just as well become a
> competing process or maybe even the dominate one for people who get
> bored but don't want to exactly die.
>   
Assuming merging is doable. You cannot just add together two neural 
networks. It works in a few special cases like the Hopfield attractor 
network since it has perfectly additive learning, but given that that 
kind of network also has a total crash in performance when it gets 
overloaded because of its additiveness, it is not a good model for an 
upload mind. If you have a typical neural network, copy it, train them 
separately to learn maps A and B on domain X and Y, and then try to 
merge the copies together there doesn't seem to be any general method to 
produce a network that has the output A on domain X and B on Y.

I would love to see some results here.

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