© 2021
In touch with the world ... at home on the High Plains
Play Live Radio
Next Up:
0:00 0:00
Available On Air Stations

Radio Readers BookByte: My Friend Al (Go Rhythm)

Wikimedia Commons

In “Homo Deus,” Yuval Noah Harari’s last chapter posits humans as algorithms.

First, we need to clarify the word “algorithm,” whose meaning has morphed to mean almost any computer program. An algorithm is a description of a computer program in plain language and enough detail that working programs in real computer languages can be written using the algorithm as a guide. Here is a very simple example from a program fragment using an if/else decision:

The Algorithm:

If the color of the object is red
run the first program
otherwise run the second program.

Below are two examples of actual program code. Notice how the actual language code differs from the algorithm.

The task described in the algorithm, as written in the BASIC language and in VBScript :

(1st line has comment at the end starting with ‘ )

If obj = “red” Then  ‘check for color attribute of red
End If

The task described in the algorithm, as written in JavaScript, PHP or “C” languages:

(1st line has comment at the end starting with // )

if(obj == “red”){    //check for color attribute of red
} else {

Harari talks about computers being far more efficient than humans. That depends on what you count in the process. Computers are brute force devices which seem quick and efficient only because they are running huge numbers of operations fast enough without showing the steps. But biological brains have shortcuts to recall information and get to answers, that computers don’t have.

Harari talks about AI programs (he says algorithms) will be wiser than humans, will know us better than we know ourselves and that we will let algorithms run our lives.

Well, one of the oldest computer aphorisms is “Garbage In, Garbage Out” (“GIGO”). An AI “learns” by finding and cataloging apparent relationships in the training data which is thrown at it. A computer program neither knows nor cares whether it is getting good data in or “garbage in,” only that it is looking for matches within the data fed to it. It cannot “decide” to look outside the provided data.

Facebook routines which feed news to you “learn” your “preferences.” Those “preferences” are an assumption by their programmers that what you click on or search for represents what you prefer to see. Each click means Facebook will look for similar items to deliver. It becomes a self-referencing feedback loop. As the number of clicks increases, the range of materials narrows.

Job offerings are now largely automated. I submitted my resume to various job sites listing programmer, reporter, photographer and videographer. I shoot mostly dance and taught for five years a course at the UMKC conservatory’s dance division, to dancers, for 3D animation of their choreography so that added college instructor to the list, with my partner, Nicole English. And I was in the Air Force from 1968-1972 as a geodetic computer and geodetic surveyor. Those are the items I fed numerous job sites.

Any human, including the newest of hires, would recommend the jobs listed. But the job sites I submitted resumes to used AI bots, not people, to parse and digest my resumés.

Right away several sites declared me a professor of ballet. I called the dance division at UMKC and told them I was ready to take up my responsibilities. I must be a professor of ballet, I said, because my friend, Al (as in Al Go Rhythm), had proclaimed me a professor there. We all laughed, but it is telling about the limits of AI and is a warning about the effectiveness of turning our lives over to algorithms.

Over the last several years I’ve seen some pretty wild job recommendations for me, frequently, from dental hygienist to truck mechanic to over-the-road driver, to nurse, to apartment handyman, to CAREGiver (sic), cook and, the other week, cardiothoracic surgeon at Hays Med in Hays, Kansas.

It would seem there is just no end to my talents. (Thanks, Al.)

For me, almost all the listings are waste, but that waste is largely out of sight of the heads of companies listing jobs. They see it as cheaper and easier than using humans, an irony for web sites whose stated purpose is connecting people and jobs.