THE TROUBLE WITH ALGORITHMS

Computer dynamics — not just the vocabulary — are turning up in the recent series plots that deal with criminals:  “NCIS”, “House of Cards.”  The first time I helped design an algorithm was in 1997 as a clerical specialist for the Bureau of Buildings in Portland, OR, which was being severely criticized for tolerating slum landlords, meaning capitalists who bought up old deteriorated property and rented to low-income people at low prices in exchange for them not complaining about the lack of maintenance.  

This is the apartment building.  Considerably upgraded.

(The old apartment building where I was living was a good example.  One corner of the top floor was open to the sky.  The occupant was a black man with AIDS, a marginally paid activist.  But the inspector I pressed into visiting was very nervous about retribution from the well-connected owner.  Neither did the occupant like me rocking the boat.  Indeed the renter left quietly.) 

We used our access to tax roll assessment, Section 8, taxes actually paid, infrastructure, evictions, small claims court, and so on, to list and rank property owners, even researching shell disguises.  It caused a HUGE outcry from the slum landlords, but praise from a crusading Presbyterian minister.

Now I just complain about Netflix, whose formula has no relationship to what I want to watch.  On medium.com the complaint is that the formula turns into a boring cookie cutter of writing that never leaves the little circle that fits the formula.

A little detour:  I’ve been reading about sexual identity, writing that is part of a long and wide conversation symbolized by the rainbow and a string of letters:  LGBTQ.  Lesbian/Gay/Bi/
Trans/Queer  and some add a question mark.  But some people are situational (one orientation in this setting, something different in another setting)— some cultures are situational.  

It looks like mom-and-pop dyadic relationships are on their way out.  But as long as meiosis (combination of two genomes) is the basis of sexuality, the biology will persist.  It’s not the ONLY basis, but if the other ways to relate don’t crush it, we’re okay -- family pair bonds will still form.  They’re biological.

Heterosexual fertile and committed relationships create a triangle which is the basis of the traditional family in our culture.  The rest can vary.  So the triangle of genetic “family” can be seen as the top of a relationship pyramid, the peak to which we aspire, or as the foundation, the basis of the whole pyramid from which the rest derives.  Hmmm.  Sex, economics, war, and inheritance push and pull and sometimes demolish everything, top down or bottom up.

I made a speculative little progression about relationships that goes like this:

1.  Awareness:  you register the existence of another person.

2.  Affinity:  the pair of you recognize interests and practices in common.

(This is what the algorithms are trying to figure out.)

3.  Attunement:  you may not agree but you enjoy the exploration.

4.  Attachment:  belonging and common ground.

5.  Physical interaction: sex.

6.  Bonding: legal binding of the stable emotional and economic relationship.  (Some cultures start at this end.  In that case, they still tend to go through the remaining steps.  Awareness may be late to develop.)

Today’s acceptance of jumping directly to coitus without the previous steps — let alone the practical supportive context of family, friends and culture — seems quite like turning the succession on its head and then cutting the head off before it even gets to attachment — preventing both fulfilling sex and commitment.  Family is removed, both that of origin and that produced.

Returning to the idea of algorithm, trouble arises in two ways.  One is the whole assumption that the essence of someone can be captured in a formula and the other is the use of such formulas in business to cluster marketing, thus never marketing to the between-formula or unrecognized folks.  We are a middle-class society, the middle-class is based on trade and transaction, but narrowing the categories of marketing clusters by using formulas makes them shallow and repetitious.  Newly forming opportunities are lost because they don’t fit.  In the long run this damages business.  

But then, thankfully, new algorithms and services arise.  If they keep up, that’s good.  If they don’t, profits slide.  If they get too far out ahead of the actual processes in the culture, their services can be attacked.  I’m thinking about sex.  Even in that context, above the level of street survival, a sex provider develops a customer base using the awareness, affinities, attunements, that can support sex.  (It’s possible for the relationship to be one of challenge as much as fulfillment.)

The balancing edge of a customer base is economical providing of service without shorting the customer so much that they go elsewhere.  Now I’m thinking about churches and stores.  This is the major mistake our government is making: not providing the services needed on grounds of economy.  Since the population can’t very well go to an alternative government in any numbers, their pushback is in sub-groups, some of them covert.  They grow. 

[Incidentally, what does it mean that Scalia was so absorbed in secretive sub-groups based on wealth, status, and old European mythology?  (which is an algorithm).  The “black ops” of the Supreme Court?  (I’m referring to the Order of St. Hubertus and Opus Dei.)]

Survival of individuals and groups depend on each other.  If too many individuals don’t survive, there is no group.  If the group doesn’t survive, it’s members are gone so it doesn’t exist.  The most basic strategy of survival is joining with a partner without gender mattering, but sex and the common interest in raising the resulting children (or other happy results), has been a strong human binding force.

I think about the phenomenon of war or epidemic when buddies (“the man on my left”) form strong pair bonds with each other, only to have them torn apart by death.  The common interest of their specific survival causes attachment to make them care for each other physically.  One might think that men linked only by sex when the AIDS epidemic hit would scatter and some did.  But many stayed with partners and nursed them through their deaths.  They became changed men, perhaps more worthy men.  Is there an algorithm for that?  Can you merchandize compassion?

Sure.  All emotions are marketable.  Esp. the negative ones (fear, anger), but the “nice” ones will do, esp. if they’re only theoretical, stigmatizing and extended from a distance to justify one’s own decisions and affinities.  But in the end these are all doomed unless they have a universal dimension. 

Grieving is universal.

Years ago I read “all comparisons are odious,” and since all algorithms are comparisons, one might call them odious by nature.  When I read that quote the first time, it was attributed to one of the fancy women’s colleges Back East.  But I googled it a few days ago and found that it traced back to an Islamic scholar in the 700’s.  But then just a minute ago, I googled it again and this time Cervantes got the credit, plus some quotes from Shakespeare.  We're told that Google will tell us what it thinks will suit us.  Maybe they don't know that like Islamic scholars.

Google is notorious for being algorithmic, profit-seeking.  The merchandising classes no doubt think Cervantes and Shakespeare are easier sells.  But this is a demonstration of an odious algorithm if a person is looking for, you know, "Truth", instead of advertising.  Algorithms, though designed by humans, are applied by machines.  Industrial virtue.   They wear out.