[FRIAM] interjection from another conversation

Wed Apr 21 17:30:12 EDT 2021

I will have to rely on your fellow wizards to decode that for me. 

In the meantime, I have only to say, 

’Twas brillig, and the slithy toves 

      Did gyre and gimble in the wabe: 

All mimsy were the borogoves, 

      And the mome raths outgrabe.

“and that goes for your goddamn cat, too!”

Nick 

Nick Thompson

 <mailto:ThompNickSon2 at gmail.com> ThompNickSon2 at gmail.com

 <https://wordpress.clarku.edu/nthompson/> https://wordpress.clarku.edu/nthompson/

From: Friam <friam-bounces at redfish.com> On Behalf Of Marcus Daniels
Sent: Wednesday, April 21, 2021 2:47 PM
To: The Friday Morning Applied Complexity Coffee Group <friam at redfish.com>
Subject: Re: [FRIAM] interjection from another conversation

The `algorithms’ that don’t work die out or else use up all the oxygen and die out, etc.   In my backyard (and most backyards here) oxalis spreads rapidly.   The `program’ doesn’t crash because there is so much `compute’ to draw upon.    A NAND gate can be embedded in a cellular automata like an Ising spin system, and all of classical compute can be built on that, and so I would think any AI we could envision.

From: Friam <friam-bounces at redfish.com <mailto:friam-bounces at redfish.com> > On Behalf Of thompnickson2 at gmail.com <mailto:thompnickson2 at gmail.com> 
Sent: Wednesday, April 21, 2021 1:31 PM
To: 'The Friday Morning Applied Complexity Coffee Group' <friam at redfish.com <mailto:friam at redfish.com> >
Subject: Re: [FRIAM] interjection from another conversation

I will answer your “no free lunch” with my, “You get’s no bread with one meat ball.”

I am currently dazzled by the interdependency of things.  To a person so dazzled, the mystery is not what natural selection (or artificial selection, for that matter) can accomplish.  The mystery is how does natural selection happen in the first place.  Something has to scaffold selection before it even possible. The mystery is modularity. If I understand you metaphor correctly, then I am led to ask, what guarantees that any of the proliferated algorithms don’t just crash the program. 

n

Nick Thompson

ThompNickSon2 at gmail.com <mailto:ThompNickSon2 at gmail.com> 

https://wordpress.clarku.edu/nthompson/

From: Friam <friam-bounces at redfish.com <mailto:friam-bounces at redfish.com> > On Behalf Of Marcus Daniels
Sent: Wednesday, April 21, 2021 2:23 PM
To: The Friday Morning Applied Complexity Coffee Group <friam at redfish.com <mailto:friam at redfish.com> >
Subject: Re: [FRIAM] interjection from another conversation

Must a developmental system void large variability?    One could imagine training an autoencoder with different underlying algorithms and/or hyperparameters and then using them in a opportunistic way.   If there is No Free Lunch, then the natural thing to do is to proliferate algorithms.

From: Friam <friam-bounces at redfish.com <mailto:friam-bounces at redfish.com> > On Behalf Of thompnickson2 at gmail.com <mailto:thompnickson2 at gmail.com> 
Sent: Wednesday, April 21, 2021 1:06 PM
To: 'The Friday Morning Applied Complexity Coffee Group' <friam at redfish.com <mailto:friam at redfish.com> >
Subject: Re: [FRIAM] interjection from another conversation

Marcus wrote :  Is that right?

Nick responds: is which right?

N 

Nick Thompson

ThompNickSon2 at gmail.com <mailto:ThompNickSon2 at gmail.com> 

https://wordpress.clarku.edu/nthompson/

From: Friam <friam-bounces at redfish.com <mailto:friam-bounces at redfish.com> > On Behalf Of Marcus Daniels
Sent: Wednesday, April 21, 2021 1:30 PM
To: The Friday Morning Applied Complexity Coffee Group <friam at redfish.com <mailto:friam at redfish.com> >
Subject: Re: [FRIAM] interjection from another conversation

Is that right?   I think sometimes I am the most creative when I am cornered.   When a particular problem comes into tight focus that *must* be solved.

I wonder if skill and modulated attention is really that valuable.   And are there really 7.8 billion unique ways to learn?   Or might some set of say 100 future AI algorithms be better (more creative) if scaled-up to have open-ended, precise memory?   Does skill learning really amount to that much, or is it just hard for humans, and a good AI could make all that suffering go away?

From: Friam <friam-bounces at redfish.com <mailto:friam-bounces at redfish.com> > On Behalf Of thompnickson2 at gmail.com <mailto:thompnickson2 at gmail.com> 
Sent: Wednesday, April 21, 2021 10:02 AM
To: 'The Friday Morning Applied Complexity Coffee Group' <friam at redfish.com <mailto:friam at redfish.com> >
Subject: Re: [FRIAM] interjection from another conversation

Hi Dave, 

And Hi Jenny!

What is the metaphor the writer has in mind when he speaks of “evolution”.  Evolution-by-natural-selection works by decreasing the variability offered up by mutation.  Evolution-by- phylogenesis may increase variability or decrease it, depending on the demands of the environment.  Stable environments offer the opportunity for phylogenesis; variable environments suppress it.  Developmental systems, of course, don’t tolerate massive random variability.  For a variant to survive most things have to be held constant.   The model Gabriel has in mind here seems to be like that of the Preconscious advanced many years ago by Lawrence Kubie in his Neurotic Distortion of the Creative Process which views craziness (eg Van Gogh) as a limitation of creativity, not as a maker thereof.  (Think of how brilliant Van Gogh would have been if he HADN’T been crazy) True creativity (says Kubie) arises from the PRE-conscious, e.g., daydreaming,  in which thoughts get scrambled into novel combinations.  But even Kubie-variation would be too great for a developmental system.  In general, I think valorizing variability obscures the fact that in any developmental or evolutionary system MODERATE levels of variability are essential, and the essential question is how do we balance the future-seeking power of variation against the stability necessary for that variation to find expression. 

Nick 

Nick Thompson

ThompNickSon2 at gmail.com <mailto:ThompNickSon2 at gmail.com> 

https://wordpress.clarku.edu/nthompson/

From: Friam <friam-bounces at redfish.com <mailto:friam-bounces at redfish.com> > On Behalf Of Prof David West
Sent: Wednesday, April 21, 2021 8:52 AM
To: friam at redfish.com <mailto:friam at redfish.com> 
Subject: [FRIAM] interjection from another conversation

A comment by Richard Gabriel that I stole from another conversation (the only FRIAMMER in that conversation is Jenny Quillien who used to attend the mother church before moving to the Netherlands).

"I think it adds a dimension to many of the evolution conversations we have had the past couple of years.

I recommend reading either “Why Greatness Cannot be Planned: The Myth of the Objective” by Lehman and Stanley or its more technical foundational paper "Abandoning Objectives: Evolution through the Search for Novelty Alone,” by the same authors. In this book and paper they argue that natural and artificial evolution are (better thought of as being) based on (or implemented as) novelty seeking with survival as a (boring) constraint. The main step of evolution, they say, is to produce diversity / novelty so that the new creation creates a new niche: the new mutations don’t compete with others for resources, they exploit different resources. For example, when exaptation created the first flyers, all of a sudden a whole raft of predators became irrelevant. 

So evolution is looking for new ways to live instead of better ways to live. Within an established species, it might be that evolution as optimization creates incremental improvements.

If evolution is based on creating novelty, then after all the simple (different) ways to live have been tried, the only direction to go is complexity."

davew

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://redfish.com/pipermail/friam_redfish.com/attachments/20210421/61391954/attachment.html>