This week’s New Yorker has an article describing Gordon Bell’s MyLifeBits project. I’ve heard about this project for years, and I’ve never understood what the point is. Collecting all of those pictures and articles and emails about one’s life just creates an overwhelming mass of data that can’t be processed effectively. It’s like the shoeboxes (or now file folders) of photos that we all have that we never look at because they’re not organized and we can’t find the one we want.
The dumbest thing I read in the article was their failed attempt to write software to do “auto-storytelling”.
“What we’re storing, really, is snippets and scenes and episodes,” he says. “A movie tells a story, and we’re not doing that.” What Bell and the others imagine, ultimately, is a function by which a computer assembles a person’s autobiography, which he or she authorizes and passes on.
What they’re talking about is creating the self story. But no computer can do that. They are operating under the misguided belief that facts alone tell the story of one’s life. I mention it in that post, but history is so much more than a recitation of facts. It’s the assembling of facts retroactively into a pattern that allows us to make sense of the facts, to tell the story.
We don’t know what the facts mean to us until we look back later and find out what happened – then we know which facts mattered. Let’s imagine that I’m out at the bar having a drink. I see a cute girl, talk to her for a bit, and get her number (indulge my fantasy here for a second). Is this fact significant in my later life? We can’t tell at this point. Maybe I get dissed after one date. Maybe we go on to have a long-term relationship. The fact of me meeting her in a bar stays the same, but the story changes depending on what happens.
The whole point of telling a story is that it lets us understand the underlying patterns that drive events. Stories are our way of making sense of the world. We feel uncomfortable when events happen that are beyond our understanding. As soon as we can come up with a story that explains what happened (even if it doesn’t fit the facts exactly), we feel better. “He dumped me because he wasn’t ready for a real commitment.” “She dumped me because she thought I didn’t make enough money.” “I got fired because my boss had it in for me.” The story doesn’t even have to make sense to somebody else – the story is the way of coping with the unexpected sequence of events.
Another problem I have with the MyLifeBits project is that it’s inefficient and unhelpful. I don’t want raw mass data dumps of my life. If I were a manager, and I asked my team to get me the sales numbers, I would be furious if they just dumped spreadsheets on me. They need to analyze the data, figure out the trends, and summarize it for me in what is essentially the story of sales.
I’m reading the book Stumbling on Happiness, by Daniel Gilbert. Gilbert cites research saying that our brain does that analysis for us. Our brains don’t remember all the details of our lives – it remembers key points, the bits that differed from the routine, the outline of the story. When we are reconstructing what happened, our brain fills in all sorts of details that may not be strictly factual, but that were irrelevant to the story so they weren’t stored. The brain stores the high-level management report, not the original raw data.
There are risks associated with that approach. As a trained scientist, I understand the importance of keeping the raw data around so that I can go back and re-analyze it later. But re-analysis takes time and energy, so it makes sense that the brain evolved to optimize for efficiency. One of Gilbert’s points is that we are deluding ourselves if we think we remember the raw sensory data, and we need to take into account the high-level nature of our memories. But I’ll get to that in a separate book review.
Scientists and engineers love raw data. It has a primal quality to it, a sense of endless possibility – who knows what patterns will emerge? But raw data is incomplete. It doesn’t mean anything. It needs us to make sense of it, to find the patterns, so that we can use those patterns to make future decisions. Without stories to bind it together, raw data is useless.
I’d much rather have this blog than MyLifeBits. This blog is where I take all of that raw data from my life and simmer it down into key points that I want to remember. This is where I record the stories and thoughts and ideas that help me make sense of the world, that I can use to make decisions. It may be a compressed representation (the backup is only 4MB), but it’s far more helpful to me than the reams of data coming out of MyLifeBits.