[MURG] new Road Map draft

[cat13 at illrepute.org]

"king-yin yan" <y.k.y at lycos.com> writes:

> From: cat13 at illrepute.org
> >Pardon - when I said "bus" I meant "multiplexing bus", which
> >would take the myriad wires within the skull and turn the information
> >into myriad signals going down a smaller number of wires leaving the
> >skull.
> 
> Good point =). That solves one problem. But the skull probably
> needs to be exposed a lot to give access to extensive interfacing.

For the surgery, yes.  After the surgery, only a small hole would need to
be exposed.
See:
http://a1112.g.akamai.net/7/1112/492/12h/static.wired.com/wired/archive/10.09/images/FF.Sight.SH.29337.r2_f.jpg
from:
http://www.wired.com/wired/archive/10.09/vision.html

> Another problem is whether it would be very expensive to
> sustain the body for a long time.

Why would it be more expensive than it is now?

> >Different areas of the brain work differently, no?  Sure, they all
> >have neurons and synapses and encode information in those low-level
> >pieces, but the ways in which they encode information and the ways
> >in which they function vary from area to area - no?
> >So, tin order to create a whole-brain interface, we'd need lots of
> >separate interfaces to communicate with each of those different
> >areas, each of which would have to be developed separately.
> >Sure, connecting electrodes to each of the areas might use the same
> >process, but interpreting the information from those electrodes
> >would differ.
> 
> I don't think we'll develop different interfaces separately.

We already are.  For example, the cortical interface I mentioned above is
being developed separately from the artificial hippocampus which has been
discussed previously:
http://www.newscientist.com/news/news.jsp?id=ns99993488
which, in turn, is being developed separately from cochlear interfaces.
Admittedly, the cochlear interface isn't connected to the brain itself,
but it still talks directly to nerves, which is the issue at hand.

> Most
> likely it will be assisted by computerized automation. Also, we
> can replace each deep-brain structure bit-by-bit gradually. All
> we need is some generic NN emulation software. This is the
> basic tenet of mind uploading -- that we don't need to fully
> understand the brain.

This brings up an interesting question.  I agree that, when discussing
uploading in general, we don't necessarily require a full understanding
of the brain.  However, is that still the case in gradual uploading?
It seems that, unless we can develop an artificial neuron capable of
synaptogenesis, we won't be mindlessly replacing neurons but
intelligently replacing functional areas comprised of multiple neurons,
and that we'd need to understand how those areas function before being
able to replace them.

> One thing I want to point out is that our
> ability to replace brain structures will most likely occur
> concurrently rather than step-by-step one structure after
> another. So we have to prepare for this scenario.

Until I saw the aforementioned artificial hippocampus, I would not have
believed you; now, however, I think that you have a good point.

> ALSO there is the possibility that we'll get uploaded first
> and then download other modules as they're available. For
> example "Vessopressin receptor subtype patch version 1.2.3"
> and so on.

Yes, that's pretty likely.  Not to be too self-referential, but we've
already seen it happen in the aforementioned cochlear implants.

> One problem is when we get to the point of experimenting
> with the first human / primate uploads, those uploaded
> entities may have some "aberrations" from the biological
> brain. Do we shut them down if things get out of control,
> or what?

Interesting.
If gradual uploading happens, then we wouldn't need to shut them down;
we could just make adjustments to the newest prosthesis.