[FRIAM] AI

[steve smith]

Pieter  wrote:
> Just one thought to toss into the mix: humans didn’t evolve to do 
> astrophysics, drive Ferraris, or detect sarcasm on Twitter....
The human *genome* did not evolve *specifically* to do all these things, 
however at some point, our facility for symbolic language and 
abstraction *did* evolve to support and enhance those more fundamental  
needs.  It is on top of that symbolic abstraction where *culture* began 
to evolve and this is where our ability to do 
astrophysics/ferraris/sarcasm emerges.   And it might be apt to notice 
that AI is much closer to driving ferraris, doing astrophysics and 
detecting (or generating) sarcasm that it is at being an effective 
hunter-gatherer on the tundra/savanna/jungle/boreal-forest.
> Now, if we set out to design a robot to function in today’s 
> environments — say, hospitals, homes, or corporate boardrooms — we’re 
> working with a very different set of goals. ...
If our goal is to replace *one more* set of skills or abilities, then it 
is correct that we don't "need all that".   It does seem that a hallmark 
of modern human activity (neolithic forward?) has been to replace 
ourselves, one feature at a time.   Lithics to replace (enhance) our 
teeth/claws, cooking to replace (enhance) our digestive abilities, 
animal husbandry to replace/enhance our brute-labor and translate 
low-grade photosynthesis to human-digestables (turning grass and leaves 
into milk, meat, eggs, blood).   Wheels and levers and sails and hulls 
and millstones and kilns and forges and hammers and anvils and looms and 
... and rocket-ships and a dyson-sphere-of-computronium  all represent a 
scaffolding of escalating replacements for the things we did with our 
own biochemistry (from hair and nails to lymphocytes and neurons)...
> So even though a robot might never replicate the full sensory richness 
> or biochemical subtlety of the human body, it may not need to....Think 
> of calculators: they’re completely clueless about context, but they’ll 
> beat any of us in a mental arithmetic race, every time.
even a well-handled abacus or slip-stick can do this of course, but yes
> I wouldn’t bet on a human-equivalent robot appearing next year — but 
> ten years? Maybe. Especially if we stop trying to replicate every 
> biological quirk and instead design for function. And when I say 
> “function,” I mean not just doing what a human can do, but doing what 
> the job needs — which is often a very different thing.
The point, I would claim is that we aspire (with AI, starting with 
Golems and Frankenstein's monster and enlightenment age humanoid 
mechanicals) to replace our "generalist" abilities.   Many domestic 
animals have been adopted/bred/trained to be "generalists" around a 
large portion of our needs, whether it be converting simple carbs into 
high grade fats/proteins for us, or hauling burdens, or even (think 
elephants) delicately manipulating things way to heavy for us.   Some 
even have diverged from direct, immediate response to our needs to more 
abstract needs we have (e.g.  show animals whose functional abilities 
might never be exercised outside of the training/show-rooms).
> Take Demis Hassabis’ current project: trying to simulate a single 
> biological cell to improve drug discovery. Sounds simple — it’s just 
> one cell — but it’s turning out to be a mammoth challenge. Meanwhile, 
> a useful robot doesn’t need even one biological cell. It just needs 
> actuators, sensors, and some reasonably clever code. This illustrates 
> a broader point: biological systems are complex because evolution took 
> the long road. Engineering can often take a shortcut.

And while I am very much a fan of engineering, I do believe Evolution to 
be more than "less efficient Engineering".   AI/ML has been an effort in 
*reverse-engineering* our greatest cognitive abilities, up until the 
recent generations of "model-less modeling" that in fact seems to be to 
try to reverse-engineer evolution (exemplified by genetic algorithms and 
neural nets and ???)

>
> So yes, the human body is a marvel — a product of billions of years of 
> trial and error. But that doesn’t mean it’s the most efficient 
> solution for every task. It’s just the one that happened to work well 
> enough to keep our ancestors from being eaten.
>
> After all, birds fly beautifully. But when we wanted to fly, we didn’t 
> grow feathers. We built jets.

In a certain sense yes, but not particularly in virtually every other 
sense?   I have had lucid dreams about flying since I was very young, 
spanning many forms from sailing/soaring to flapping, to pumping my arms 
as with a hydraulic jack to telekineticing, swinging from vines or 
spider-man-shot webs, and more recently to following stellar/planetary 
geodesics with my intention, spreading my electromagnetic pinfeathers in 
the solar/wind/magnetic fields of the 
sun/planets/moons/belts/rings/clouds.   But no jets... or any mechanical 
contrivance really...  even icarus/daedelus wings-o-wax...

<aviation tangent>

    While I have flown a (vintage) light plane, even through minor
    aerobatics, and ridden hundreds? of thousands of miles in commercial
    jets, none of it really matches the experiences I
    *aspire*/*imaginate* to have in those dreams...  I can't (until
    DaveW or maybe Glen turns me on to the right entheogen) project my
    spirit into the Ravens frolicing on the canyon edge or the bats
    echolocating their way through a flux of airborne foodstuffs.

    But yes, Jets, hypersonics, space rockets, interplanetary bussard
    ramjets, ultra-lights, gyrocopters, etc.   They are marvels and each
    in their own way more "capable" than any given bird/species and
    there are probably unpiloted winged vehicles which will rival the
    Wandering Albatross reputed to spend 95% of their time in the air
    and expending order 2x their basal metabolic rate.   See Gossamer
    Condor/Albatross/Penguin and note that Aerodynamic Engineer on the
    first two projects, Peter LIssaman hung at SFx with us for a while?

        https://en.wikipedia.org/wiki/List_of_human-powered_aircraft

    How soon will we have a Lissaman/MacReady equivalent AI (throw in
    pedal-pusher Bryan Allen for good measure)?  Or leonardo DaV or
    Archimedes, or... ?   Can they exist/arise/emerge without the larger
    culture that spawned them?   Maybe we can identify to an AI their
    achievements and reverse engineer their solutions, but can we frame
    the context that lead them to pursue those challenges.

    I *suspect* that if given the opportunity/motivation that suites of
    AI/ML agents might well develop minimalist human-scaled prosthetics
    so that I (my grandchildren) can literally have those experiences
    direct.   Some proprioception/motion-platform/haptics added to
    visual/auditory synthetic sensoria and even I might be soaring
    virtually over the surface of mars or through the braided rings of
    Saturn as-in-my dreams?

</tangent>

And in the immortal /Tears in Rain/ words of Roy Batty:

    /*"I've seen things you people wouldn't believe.*
    Attack ships on fire off the shoulder of Orion.
    I watched C-beams glitter in the dark near the Tannhäuser Gate./

    /*All those moments will be lost in time, like tears in rain.*/

    /*Time to die."*/

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20250621/6e40c25c/attachment.html>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_0xD5BAF94F88AFFA63.asc
Type: application/pgp-keys
Size: 3118 bytes
Desc: OpenPGP public key
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20250621/6e40c25c/attachment.bin>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: OpenPGP_signature.asc
Type: application/pgp-signature
Size: 840 bytes
Desc: OpenPGP digital signature
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20250621/6e40c25c/attachment.sig>