Alternatives
When Amazon changed the encryption system AGAIN on the Kindle I put it in the closet and switched to a Kobo. Same idea, but backed by Rakuten in Japan. I get all my ebooks from them now.
338 publicly visible posts • joined 31 Dec 2018
I kind of liked the Good Old Days where the computers were in a locked room with a raised floor, Significant air cond, and a specialized crew of acolytes to take care of them. Luckily for me, I was one of the anointed software people who knew the combination to the door lock.
When I started university an EE degree took 5 years and included a required course on thermodynamics. Reports from upperclassmen was that this course was a real bear as the Prof was more focused on the calculus than on getting you to understand heat transfer intuitively. Fortunately for me, halfway thru they revised the curriculum to reduce it to a more normal 4 years but this required you to choose a sub-specialty. Mine was "computer engineering" (heavy on software, less on building the things) and did NOT require the Thermo course. I am still mindful of cooling and always install the "optional" heat sinks and things even on little Single Board Computers.
Human programmers are very lax at putting comments in the code to explain, not WHAT it is doing, but WHY it is doing it, and WHEN it will be doing it. I wonder if AI puts in useful comments. I think of code comments as messages to a future me who might need to debug it years from now.
In the 1973 movie "The Paper Chase", about attending law school, there is one character, Kevin Brooks, who brags about having "a photographic memory". He can recall details about every past case from the law books. But as the Professor points out, that is of no use without the ability to reason and to work out the hypothetical cases that are the foundation of learning the law. Generative AI is like Kevin Brooks, with the added feature of being unable to say "I don't know," and making stuff up instead.
IBM knew about this risk back in the early 1960s, which is why the emergency power off switch on all IBM 360 and 370 computers required you to PULL on a red knob at the very top of the console panel.
But other people at IBM put the LOAD button (which means "reboot" in modern lingo) just a few inches above the desktop where you might bump it, hit it with you coffee cup, or push a telephone against it.
1. It will tell you that whatever you suggest is a really good idea and offer help to accomplish it. Managers will love this.
2. AI will make you average. Below average people will think it is wonderful. (This is 50% of the population, perhaps more among managers.) People who are already above average in reasoning will be brought down to that level or realize that the whole thing is a waste of time.
3. Even when using AI just for research, it will make things up.
The comments (embedded, within the code, not in separate documents or header file) need to explain the PURPOSE of the code - not just what it does, WHY it does it this particular way, and under what conditions it can be executed. Those are concepts that can not be expressed in any programming language I know.
Without that information, the code is un-reviewable and un-maintainable, EVEN LATER BY THE PEOPLE WHO WROTE IT.
I heard a nice description of AI making your work average. So, if you were below average to start with, it is an improvement and you are impressed. Count most Managers in this group. But if you were ABOVE average, it drags you down as you spend more time finding the mistakes. Remember what George Carlin pointed out: you know how stupid the average person is? Well half the population is worse than that.
My worst experience was a very old Epson printer that had an ink hose come loose inside the case. What a mess. I had to buy a new one. (The new Epsons are much better.)
The riskiest think I ever did was replace the ink ribbon on an IBM 1403 printer. They put those plastic gloves in the box for a reason...
China has changed CONSIDERABLY in the last 20 years, getting rid of the pollution in cities and lifting about 850 million people out of extreme poverty (according to the World Bank). Things move fast there. Healthcare is widely available and very inexpensive.
"The Art of Computer Programming" by Donald Knuth (1968 plus additions) was my bible.
I did once give AI a simple programming problem that should have taken about 8 lines of code. It made a fundamental error. No, a twos-complement addition is NOT the same as a boolean OR, jeez.
I was asked to review Volume I of set of book written by our assigned technical writer that explained a new programming language which was to be the new standard for system software development in the company. I started into this project and immediately started finding errors, ranging from simple grammatical mistakes, to clumsy presentations, to outright technical errors. Soon I realized that a few scribbled notes would not be enough to convince management how bad this was so I decided to slog through the entire book (about 80 pages) and write down every little thing that was wrong. My finished letter was 15 pages, single spaced. My final summary was "get somebody else to write the book; preferably somebody with programming skills."
My manager thanked me for the useful and detailed feedback and asked me to do a review of Volume II. I refused to waste my time, as they had not understood the point of my effort. Complex programming languages (which this certainly was) require a certain mathematical precision in comprehension that the assigned writer was not up to. For example, compare "Each statement is terminated by a semicolon" (what he wrote) with "A series of EXPRESSIONS are SEPARATED by semicolons" (how it actually worked). Notice the confusion about what do you put after the last expression in a sequence. With no semicolon, the value of the sequence is the value of the final expression. If you DO put a semicolon there, the value of the sequence is zero.
Lots of AI Slop on YouTube as well, with generic images taken from free libraries, rather obvious narration pasted together from news stories, and a voice that mispronounces some of the words and subtitles that put the wrong word in. I notice that on the Chinese social media platform xiaohongshu (known as "Redbook" to Westerners) AI-generated content is sometimes flagged as such.
About 30 years ago when I was working for Oracle, the corporate President at the time came to our far-flung office to talk about strategy. He was quite open and up front about it: it was entice people in with features and then lock them in so they couldn't leave. I guess followed by holding them upside down and shaking all the coins out of their pockets.
I was using the Bliss-16 compiler for a project. The Bliss compiler was really good at optimization for memory usage. I once saw it generate the instruction "MOV @PC,4(R2)". That stores a literal 4 at 4 bytes beyond where R2 points but takes one word less than the more straightforward "MOV #4,4(R2)".
There is a famous book on this titled "The Peter Principle". It's premise is that people get promoted because they were competent at their previous assignments. This continues until they are promoted into a position for which they are INcompetent. Since promotions require demonstrated competence, they then stay there, being incompetent.