[ExI] Uploading

[Alan Grimes]

chrome://messenger/locale/messengercompose/composeMsgs.properties:
> Alan Grimes <agrimes at speakeasy.net> wrote:
> 
>> it might be
>> possible to build a hybrid mind architecture with a very
>> non-zero chance
>> of success and with benefits far beyond what anyone has
>> ever promised on behalf of uploading. 

> ...

>> Are there any real transhumanists on this list even
>> slightly interested
>> in becoming a million times smarter than they are right
>> now?


> Now you're confusing me.

No, it's your fundamental nature to be confused. I can't help that. =(

> First, you appear to be vehemently against the very idea 
> of uploading (I'm still not sure why), then you throw in 
> the concept of "a hybrid mind architecture" using 
> neural interfacing?

> You've just described one uploading scenario!

No.

I haven't.

Uploading is very strictly defined as a destructive scanning of the
brain followed by a simulation of that exact brain pattern (including
all of its defects) in a computer. This procedure is discussed in great
depth in the "brain emulation roadmap" by Anders Sandberg, which I
consider to be the definitive work on the subject. In *ALL* discussions
of the subject, without exception, since the beginning, the destructive
procedure has been assumed as the working model. In 90% of the other
papers on the subject, the destructive approach is again assumed. In the
recent discussions on this list, there have been no shortage of people
defending the destructive approach, and accusing me of being
delusional/stupid for thinking otherwise.  (Due recognition of people
who proposed other procedures such as reversable uploading.)

Nobody, anywhere, thinks of a situation involving an AI substrate and a
neural interface when someone mentions uploading. I invite you to spend
the entire weekend trying to find a serious proposal by a prominent
author for any kind of hybrid intelligence that is not intended only as
a stop-gap for full destructive brain uploading.

Uploading == destructive brain uploading with full brain emulation and
*NOTHING* else.

What you are doing is attempting to co-opt me and the movement I want to
create to provide an alternative to uploading -- and that pisses me off,
Ben. =|

> There are many (if not most) people on this list interested
> in becoming many times smarter than they are now, and the 
> first step in that process has to be some form of uploading
> into a non-biological brain.  

The non-biological part, I partially agree with.

The uploading part (see above), no. Uploading is uploading. Using a Cray
XE6 as extra gray matter is not. (OMG: 1 million core scalability!!!)

> That may turn out to be the best way to unravel the 
> architecture of our minds, and almost certainly the only 
> realistic way to radically improve them. 

I am fully in favor of using uploading as a research tool. Take the
brain of an accident victim or something and upload it for the sake of
science. I'd be willing to treat such a being as a person but it's not
the accident victim, it's a robot loosely based on what happened to be
in his head after he died.

> While our minds are still implemented in meat, there's little 
> chance of being able to improve them in any significant way.

Sure they are. If you don't insist on carrying your brain around with
you, it's perfectly reasonable that it could be expanded biologically.
Advanced CMT would also serve as a mostly biological enhancement.
Additionally one could consider merging several bio-brains together.

My point though is that you are both unimaginative and wrong.

(And it's far far worse to be unimaginative).


> Your suggestion is an interesting one in that it would 
> presumably give the uploadee some control over the 
> pace and degree of the process. Some people might want to
> stop halfway, and remain part biological at least for a while 
> (Why? Beats me! But there would surely be some).
        ^^^^^^^^^

That's cuz you're an UPLOADER!!!

> Of course, their hand might be forced eventually, 
> if life-extension tech. doesn't exist to keep the biological 
> part alive indefinitely.

And why shouldn't it?

> I imagine some kind of 'parallel running' would be used to 
> fine-tune the non-biological parts, and when bio and non-bio 
> functions are perfectly aligned, the bio can be slowly phased out.
> Or maybe the whole brain would be kept intact, and the non-bio 
> part would grow so much that it was implementing the vast 
> majority of your consciousness.  Eventually losing the 
> original brain would be of no real consequence.

Yeah. But then there is no reason to perfectly align them, let each
specialize. I would imagine having a network of mind-hosts, each hosting
k/N of my consciousness. The nodes on the network would be on a
continuous rolling upgrade cycle. -- every time a new processor came on
the market, the oldest node in the network would be retired.
Furthermore, there would be a good deal of diversity in the network, I
might maintain several chains of nodes using a broad array of computing
technologies, limited by my resources for maintaining the network.

The network would be heavily redundant. I would demand perfect
reliability from each machine and at no time will more than one node be
scheduled for downtime.

When a new node is brought on line, it would simply be plugged into its
intended role and allowed to learn-adapt into the matrix and begin
functioning as a part of the whole. When a node is retired, any useful
work it is still doing will be transferred to the newest node. No
information will be directly copied off the old machine because it is
probably crufty and inefficient. Instead the old node would re-play it's
memory and the newest will learn it and organize it in the most
efficient way, filtering out any unwanted junk. But the system as a
whole will remain on-line.

A network of this type could wipe the floor with any uploader. There is
no comparison. The networked consciousness approach is a quantum leap
over stupid, backwards, uploading. Furthermore, even when working
properly, an upload will end up being stopped and rebooted all the time,
so it would be very difficult to say it's living at all. A network such
as what I propose, could remain stable for millions of years.

I'm terrified that all these myopic, dim-witted, backwards thinking,
idiotic, suicidal, uploaders will rule the world and vastly superior
approaches will be precluded before they can be put in place. =(

-- 
DO NOT USE OBAMACARE.
DO NOT BUY OBAMACARE.
Powers are not rights.