[ExI] Purified humanity
eugen at leitl.org
Mon Oct 7 14:36:44 UTC 2013
On Mon, Oct 07, 2013 at 06:05:47AM -0700, Ben Zaiboc wrote:
> It certainly looks like a hell of a mess, and it's easy to see how inevitable it is that we get ridiculous results like our inside-out eyes, and crappy spines, and all sorts of other just-good-enough-to-sort-of-work structures and mechanisms that any self-respecting engineer would commit hara-kiri over if they'd had any hand in them.
Yeah, due to the need for systems to be embodied at each step
we get atrocities like http://scienceblogs.com/grrlscientist/2010/06/22/the-laryngeal-nerve-of-the-gir/
> It's very tempting to conclude that we need to re-design the whole mess in order to gain any sort of control over it (so that we can improve our lives, extend our capabilities, lifespans, etc.). To optimise it.
A major improvement would be ability to hald and copy state to new substrate,
or even incremental remote replication of evolving state. This makes you
substrate-indepedant. It's a big step, but given that we can halt state
now (cryopreservation or fixation/plastination) it's one that is available
to us today.
> I think that, in a sense, you're both right, and a highly-optimised biology, while working much better, would also be very brittle, and lack the flexibility of naturally-evolved systems. Maybe what we need to do is build efficient systems, then make them less-efficient again by building in lots of redundancy, back-up mechanisms and ways of doing things differently (see what I did there?), but in a way that gives us much better control over the whole thing. This probably involves understanding biology really, really well. Probably so well that by the time we do, we won't need it anymore.
I think the whole buck stops about a complex system trying to understand
its own operation. You can't destill it into something comprehensible,
yet still meaningful. Even if you can write down all the equations
governing your neural tissue, this still doesn't tell you a damn thing
why you crave grilled calamari today.
> Anyway, I do think that spaghetti evolution sucks, and there must be a better way, that can get us the benefits of comprehenisble, controllable systems /and/ resiliency.
If you ever find out an alternative, do drop us a note.
> Actually, that gives me an idea. Need to think about it for a bit..
The problem with evolutionary systems is that seed is compact, and
morphogenesis is a complex, trapdoor system. So there is no simple
way to link bit mutations in the seed to the expressed adult system.
It's pretty much like trying to invert a cryptohash.
-------------- next part --------------
A non-text attachment was scrubbed...
Size: 836 bytes
Desc: Digital signature
More information about the extropy-chat