Unfortunately, I was hit by a cold, and so I have been out of it for the past week. I ran into a dubious review of a book that I have not read — “Power and Progress” by Daron Acemoglu and Simon. I started commenting on what was discussed in that review, but decided that this concept was a stretch for an article. Since I had nothing much else I could write about, I just want to salvage a couple of points from my thinking. Most of my recent writing has been cleaning up my inflation text, also discussed.
The book blurb for “Power and Progress” starts off as:
A thousand years of history and contemporary evidence make one thing clear: progress depends on the choices we make about technology. New ways of organizing production and communication can either serve the narrow interests of an elite or become the foundation for widespread prosperity.
I will assume that this offers a big picture summary of the contents of the book (and is a view that the reviewer objected to).
The problem with this statement is that it is more accurate to state: “Societies can either serve the narrow interests of an elite, or have a foundation for widespread prosperity.” That is, we are talking about mixed economy capitalism for modern developed states, and not technology itself. Discussions of technology are invariably derailed by political economy questions. Since I have not read the book, I will not attempt to speculate on its contents. I will instead discuss what I see as generic issues with the discussion of technology.
One issue with this topic is that “technology” has been misused as a term, and has been defined in practice as being whatever Silicon Valley venture capitalists are hyping. The problem with that definition is that Silicon Valley and “technology” companies are mainly hyping scams and regulatory arbitrage that are somewhat related to software. None of this has anything to do with the applied scientific research that happens across most of the non-Arts faculties in universities. (And it is entirely possible that Music faculties have people looking into digital music technology.)
The next problem with discussions of technology is that it has been politicised, with fans of capitalism arguing that technological advances have been the result of unfettered capitalism.
The final problem is important for economics. The belief is that because there was a lot of prosperity associated with technological progress in the past, this will always hold in the future. In particular, output is determines by a production function, with an ever-rising contribution from technology. The concern is that past performance is no guarantee of future results. Replacing animal and human muscle power on farms made it possible to redeploy the vast majority of the workforce to new industrial occupations. Now that job requirements are fragmented, it is going to be much harder to find similar big wins to raise prosperity. New product churn is going to allow GDP to grow, but there is a gap between GDP statistics and national well-being.
These points come together with everyone’s favourite high technology, Artificial Intelligence. (Blockchain used to be the favourite, but Silicon Valley pivoted.) The premise is simple — most people in the developed countries spend at least some time near a computer screen as part of their job, so what if we used computing power to make computers more effective tools? A new software tool can find usage in many places. The problem is the fragmentation of work — even finding a case where thousands of workers can be replaced by a chat prompt is not going to move the needle on global growth. At the same time, it does not help that a lot of discussion of artificial intelligence can be best classified as mediocre science fiction. In terms of real world effects, artificial intelligence has made it even easier to fake news and images and possibly destroyed the usefulness of web searches — but those were trends that were well underway already.
Inflation and Banking Texts
I have been plugging along with my inflation manuscript (viruses willing). I have mainly been deleting text, so here are some examples of textual improvements: _____ and ____.
I think the manuscript overall is OK, but it still has rough edges. Rather than force a publication out, I decided to et the text marinate. My main focus is trimming the length, and adding more reference material where needed.
Since my writing output dropped, I had been doing more consulting work. Although I could continue with musings about banking, I want to move my draft articles into a single Word document. It is only at that point I will be able to see what is missing from the banking text. However, I do not want two unfinished Word manuscripts hanging over my head.
One problem with the banking text is that it probably needs a lot of balance sheet examples. Something like a balance sheet is the worst possible content for me to lay out for an e-book. Theoretically, I can just use an HTML table. The problem is that I need to manually edit the HTML, which is incompatible with my existing publication flow. Doing anything fancy also run into compatibility issues with older e-readers. The formatting awkwardness means that anything involving balance sheets have to be done last, as I do not want to waste time formatting tables that just get axed during editing.
My other concern with balance sheet entries is that they might only make sense to readers who already know how things work before hand. Unless you build up the balance sheet entry-by-entry, readers are confronted by a wall of numbers laid out in a fashion in which they may be unfamiliar with. In a textbook, professors can hope that the undergraduates will pore over a balance sheet and then be able to regurgitate its meaning on an exam. I cannot impose an exam on my readers, so information has to make sense on the first pass. (My writing does not always meet that standard, but I know that is what I should aim for.)
"Automated Information" as I like to call AI, will have about as big an impact on society as the escalator, by my estimation. The escalator is my goto example when you are talking about complex technologies designed to replace simple tasks. They can work and be nice, but they're often not cost effective, almost never maintenance and worry free, and end up being used as a niche or novelty. Similar for idea for VR and other wearables.
I am fluent in html, word, LaTeX, all the rest, but merely with floss tools I now find markdown entirely adequate, beautiful LaTeX is easy to generate automatically, as are tables, and all the rest. Only inserting PNG images with a nice background color is a few minutes of drudgery, but a bash script can mostly automate that too. Sample: https://t4gu.gitlab.io/t4gu/theory/005_spacetime_algebra/#refresher-the-spacetime-clifford-geometric-algebra (that's a hugo website with goldmark and katex).
LaTeX for book projects though is still hard to beat.