Archive

Posts Tagged ‘Gordon Bell’

Total Recall

November 1, 2012 Leave a comment
Total Recall (book)

Total Recall

Here is an interesting video that I watched recently on YouTube, in which two men who ran a research project at Microsoft to record in the computer almost every single detail of their lives.

What if you could remember everything? In this lecture, Jim Gemmell and Gordon Bell discuss their new book, Total Recall How the E-Memory Revolution will Change Everything. Bell and Gemmell will draw on their experience from the MyLifeBits project at Microsoft Research to explain the benefits that will come from an earth-shaking and inevitable increase in e-memories.

This is pretty much what I’ve thought about in terms of how the use of computers might help people augment their memories. But the approach is in some ways entirely different from the approach of Piotr Wozniak, and other users of his SuperMemo program. In the approach of Jim and Gordon, you do little but passively record with a variety of devices that you carry with you. You make a small effort to digitize, upload and tag, and then rely on the computer to store the information. There is a little in the video about the search program they wrote to help you find what you might want to look for by sifting with a variety of metadata, such as time and automatically recorded GPS location data, but the main approach seems to be just leaving the records of your experience in the computer.

The most interesting thing they said about the usefulness of digitized data was in several mentions of images coming up in a screensaver. They note that this kind of presentation makes a lot of memories more accessible, it makes photographs more enjoyable, you see your photographs and pictures of art and awards more when it comes back to you like this. I agree. That is my experience too, regarding random slideshows of my large image collections.

But it seems to me that we are missing a step. A lot of information we want, not only to store, but to have available in our own brains. In experimenting with programs like SuperMemo, and the exercises available on mybraintrainer.com the idea is to use the computer, not to replace your ability to remember, but to strengthen it, to allow you to take control of programming in what you want to have quick access to.

This is not to negate any of the points that these men make in their talk and interview or in their book. There is a lot of value to be had in keeping information, in keeping records, and photographs, and to having metadata about different parts of your life, but the difference in philosophy is important. Most of what I get out of this is that these men see humans using the computer as if it were something external to us, an appliance, a super-photo album, Piotr Wozniak uses the computer as a cybernetic component, to progressively load data from the computer into long-term storage in his brain. He uses the computer to program data into his own memory, not as an externalized system to replace it.

But in some ways the lines between these philosophies are blurring. In more recent versions of SuperMemo, Wozniak has introduced tools to increase the volume of information he can load into the program. Instead of reading from textbooks, and inputting facts worth studying into the program to lock into his memory, he now loads whole articles, and reads and processes them bit by bit. To cope with the increased volume of information he studies, not all of it yet known to be worth the effort to memorize for the long-term, he has had to revise the program’s basic algorithm, so that it is no longer based entirely on the statistical model of forgetting, but instead depends on the user’s judgment of the priority of each article or question. One result of this is that when the learning program becomes overloaded, to the point that you cannot deal with all of the information it presents to you each day, there is going to be some material forgotten, this is acceptable if that material is of a low priority. In SuperMemo these features are called incremental reading and the priority queue.

Effectively, Wozniak is relying more on the computer to store information he might find useful, and less on it as a means of training the information he has loaded on it into his brain’s long-term memory. In my opinion, there is still something to be said for the routine and the algorithm, as opposed to storage and search.

And then there is serendipity. Recent versions of SuperMemo now rely on a certain amount of randomness in their algorithm, which slightly counters the strict ordering of the priority queue as a tool to overcome the user’s priority bias. Another tool that Wozniak recommends to help combat priority bias is to run a completely random review of some of the material in your database from time to time so that you can rediscover things you were once interested in but which had sifted to the bottom. I use a program called ACDSee to view my images, and with it, I like to jump through a file at random on a full-screen view. I can only image that Jim and Gordon get a very similar effect from the images they have come up on their screen savers driven by the software they use. I often think that it would be fantastic to have an image viewer that would work like SuperMemo, showing me images, not strictly at random, but based on criteria like how I have rated them, or which I had last seen less recently.

The main idea I get out of all this is that there are many ways to use computers to augment human intelligence. I think the best approach is a cybernetic one. I want my computer to store information for me, far beyond what I could remember in my brain, but I also want it to do its best to help me keep the most important information at the forefront of my mind at all times. I want it to help me keep the most useful information in my long-term memory at all times, and I want it to prompt me with visions of my most significant experiences, randomly juxtaposing disparate ideas, fostering chance mash-ups. An electronic daydream for imagination that is beyond the merely human.

The important part of all of this is not storage capacity, or input devices or storage. It is going to be the interface, that, and, to some extent the rules for deciding what to record, what is really important.

Advertisements
%d bloggers like this: