Thursday 12/19/2024
There was no CHM volunteering today (and none until 1/6, I hear) so I had an open day. I set myself the goal of reading two things that had been sitting on my coffee table too long. One was the latest edition of Scribble and Sketch, the CH literary magazine. I wish that was online so I could link to articles in it, because there are some good ones.
The lead article was a profile of Arlene (Kwok) Sullivan. I’ve mentioned her a number of times, as the person who does piano accompaniment to various performances. She’s well over 90, but is still an excellent pianist. She had many years experience as a performer with different groups, but the striking part of the profile were pictures from when she was a teenager, and a star in local Chinatown dance bands. As a very young person she sang “Any Bonds Today” at WWII bond rallies. As a mother she formed a band with her own teenage children. She taught music in several bay area schools. There are also a couple of shocking medical tales from two different doctors. At the back, they published a piece I wrote on being jaded about holidays, which I included here back on day 4.347.
The other was the book Nexus by Yuval Harari. Joanne bought that and loaned it to me to read first. I have mixed feelings about it. Harari presents many interesting historical examples to make these important points. First, improved communications do not automatically bring improved knowledge, or more democracy or more “openness” — as optimistic tech people tend to assume. He shows how the introduction of the printing press led directly to the witchcraft panics of the 1500s, in an early example of better communications only improving the transmission of conspiracy theories. For a more current example, Facebook’s algorithms strongly aided the Rohingya massacres in Myanmar: the algorithms were designed to increase “engagement” and the algorithms “learned” that violent anti-Rohingya propoganda increased engagement. On the other hand, a modern democracy can’t function without plentiful communication, but a totalitarian state can. There’s lots of complex “this, but again, that” here, but a clear warning against letting computer algorithms be decision-makers without requiring a clear appeals system. He talks about the failings of algorithmic judicial sentencing; he doesn’t mention health insurance companies using algorithms to deny coverage, that news is a little too recent for a 2023 book. In general I think the book has a lot of important historical lessons, but could have been better organized with clearer conclusions. (Maybe he should have run it through Claude.ai…)
Codgerville 