Spreading the News

 As this kind of started as a discussion of algorithms and their future, and when they will enslave the human race for reasons we will never be able to understand, I thought I would talk about copies of information. Does redundancy add information that may affect function of the algorithms?

I was working on updating slides for one of my classes because we changed textbooks for the coming quarter. When I was ready to have the slides uploaded into the Learning Management System we use, I attached them in a series of emails to the guys who do that at my school, with a copy to the Program Director. I also sent copies to all the instructors on the other campuses of our school. Sure, they had already seen most of them, because as I finished a file, I shared it with my colleagues, asking for feedback. I didn't get much feedback. I made a few edits before the final upload on some files, without another sharing. I knew the files would be uploaded, and available in the LMS later the same day, and I also knew that I had stored copies of the in a folder on the school server, so everyone already had access, usually in more than one place.

What does that mean for the algorithms? Does spreading out so many copies create any new information? Does it increase the odds of some algorithm getting that information into its system? I can't actually imagine the extra copies making much difference to anyone unless they strain the capacity of the system somewhere. One server space I have at work is nearly full. If I store copies of all my slides each quarter, whether they have been edited or not, I would very quickly run out of space, but with no added value, because every copy would be the same. The only actual information I see from all of that is that the algorithms could keep statistics on our interactions. Maybe if I share more files with my colleagues, the algorithms will see a stronger relationship among us, and will make something out of that.

The same could be true for memes and posts on social media, I suppose. Some useful information for the algorithms could be extracted from analysis of the rate of spread, a sort of kinetics study of the flow of information. There is also the possibility that some of my colleagues edit the slides before showing their students. During lectures, I suppose students could also copy slides, and make their own edits, like taking notes on the images. Apps exist for that, and I see students making notes and drawing things on pictures they take in my labs. I wonder how much access the algorithms have to those images. I don't know what information might affect the function of an algorithm. My impression of some computer system work, such as expert systems that are based on machine learning is that no one quite knows how the systems come to their conclusions, even when the expert systems show improvement in performance over the humans they are, um, replacing. That means there is information, even useful and valuable information, in forms scientists and doctors don't see, or in the ability of computers to combine larger bodies of information, and get value from big data, as it is sometimes called. Value that is unique to the systems that use it, as mere mortal brains don't have the same capacity for data storage.

So, I don't know if including everyone on all those emails does any good for anyone, even an algorithm. I do it because it is easy and seems unlikely to cause harm. I could be contributing to the end game, but I can't see it. Unless this post means that I can.