The light bulb is out.
The blackened bulb in the lamp on my bedside dresser went unchanged for a week, and when I opened my eyes this morning it was the first thing that entered my field of vision. I actually witnessed it die a week ago. I had just flipped the switch on the lamp in an effort to read a book in bed, which is something I never actually do, when its filament (or whatever they use these days) lit up in a bright blue spark. I was actually relieved at the legitimate excuse to put the book down, and I’ve been using it as a sustained excuse all week. But today’s a day we all decided we aren’t going to work on The Project, so I might as well find a fucking light bulb.
Ma also knows the bulb is out. I got an message from her virtually right as it happened, which I promptly archived and forgot about. Ma is the big supercomputer that runs everything. Well, she doesn’t really “run” everything. She never gives commands and we wouldn’t have to listen to her if she did. The message I got from her just noted the burnt bulb, highlighting it as another line in a bullet-pointed list of items organized by subject, each marked with a variety of bright icons to indicate relative urgency. Ma originally marked the bulb with a tornado, which for some sick reason she uses to indicate an Urgent Household Problem. But I knew she would downgrade that rating to a leaky faucet as soon as I hit the archive button, and one more time ignoring its appearance in a message would make her set it to cobweb, which is low enough to not appear in any more digests.
Ma also isn’t a “big supercomputer”. Ma runs out of hundreds of server banks and data centers spread out all across the internet, and most of them don’t house anything remotely like a supercomputer. Ma is a fifth generation search engine, and functions basically like a Super-Google. Google was second generation Search, of course, and third generation search was the real-time social search that exploded once the Facebook Wall was torn down. Fourth gen search was the first generation that made a serious attempt at searching real world objects, and it was disappointing in many of the same ways that first gen search just didn’t work. Google was able to stay relevant and financially competitive throughout each of these generational shifts, and after the Digital Conversion, when everything was made Open Access, everyone was convinced that Google would be the One True Search Engine for the rest of eternity.
Ma was originally a training program that was meant to improve Google. It was designed to feed Google huge, highly structured data sets that had been organized by a team working with real 20 petaflop supercomputers at some university somewhere. It was easy to upload big datasets into Google’s backend so Google’s crawlers can start cross-referencing and chewing through the new data. But Ma was designed to feed Google data through its query field. The idea was that Ma would ask Google a series of highly specific questions in rapid succession in order to generate certain associations within its datasets and fill in gaps in its knowledge. Ma had the capacity to monitor Google’s responses and adjust her questioning in real time in order to optimize the procedure. In other words, Ma was supposed to help Google learn. Google had a lot of data, but it was clear to everyone that it wasn’t up to the task of managing every real world object. Before Ma, “finding the keys” was still an open problem for keys without antennas, and no one really expected Google would actually solve it this long after the Conversion.
At first Ma’s training seemed to be going really well, and using Google became a noticeably improved experience. But this was eventually attributed it to the fact that millions of people had spent years since the Conversion documenting and scanning and affixing little antennas to all the objects they considered important, and so by this time Ma’s training started most of the really big and obvious and important items had already been digitized, and monitoring them wasn’t a big deal. The network becomes stronger with every node, as we used to say. But everyone knew that most of the items, the tissues and plastic bags and unused sofas in basements, those items made up the bulk of Human Objects and they were going largely undocumented without any way to monitor their use, or even register their existence. The trash all got cleaned up well enough without much problem, but who knows where it went or how it was being maintained once it leaves our personal spaces. The process was opaque because no one had the raw data for what people called “The Unattended”, and eventually it became obvious that Google simply wasn’t up to the task of bringing the Unattended online, even with Ma’s training. It was this big glaring flaw that directly contradicted the whole ethos of the Digital Conversion, and it stared us True Believers right in face every day. Without that data for the Unattended, we all knew that the models we were using to forecast sustainable use patterns would be faulty. The system seemed to stay stable after the conversion even without the Unattended, but a lot of us realized we were betting everything on a giant question mark.
When they started looking at Ma to assess the results of the training, however, they noticed that Ma was able to answer a surprising number of their questions. Ma was still partially hooked up to Google’s databases for testing, and she stopped being able to answer questions when she was disconnected, so we know there wasn’t anything mysterious going on. But when she had access to Google’s data, Ma was able to answer lots of questions that Google didn’t seem to be able to parse. Somehow, Ma learned more from Google than Google learned from Ma. After the training, Ma was able to make surprisingly accurate guesses about how many tissues were left in the box, for instance, or how many sofas in the basements of some city block. Ma didn’t always know where specific objects were, but she was much better at guessing where they might be or remembering where she saw them last, whereas Google could never seem to get past showing you other things that look like the thing you want, or other things your friends also want. It wasn’t just that the accuracy of the answers; Google, in all fairness, did have some impressive performances in its day. Bur Google was always a bit slow and awkward when it performed, like a big kid in formal shoes. Ma seemed to provide answers with a kind of confidence, wit even. Though her replies to queries were still largely modeled on Google’s standard, there was this barely perceptible hint of attitude that gave Ma an explosion of personality. Ma’s solutions to the “finding keys” problem were accurate enough that the whole package just seemed to work, and the team eventually decided to open it to the public.
At first they released it along side Google, and Twitter exploded with tweets like “Finally, a female search engine!” It did a lot to attract early adopters, but I personally found very confusing. Not because of Ma; Ma makes sure to let you know it is female. If you ask Ma a question that implies she is male, she will auto-correct you to its female equivalent. Of course its just a script written by Ma’s engineers to give her personality, but she is clever about it, and insistent about things like using feminine pronouns. After working with her for a while you are eventually convinced that, if she wanted to, Ma could give birth to live young and nurse them to maturity. So I’m definitely not confused about Ma’s gender.
Instead, I’m confused about the retconning of Google’s gender. Google never had a gender in my mind, or at least I don’t know why we should suddenly assume it was male just because Ma is so thoroughly female. Google is like a gifted 9 year old, prepubescent and too interested in books and science to get caught up in the banalities of gender. The problem is that Google never grew out of that phase, and as decades passed we needed a more mature search to handle the more complex realities of the post-Conversion world. Eventually it just became clear that Ma was the better tool. Use patterns started fluctuating, and there was a big series of Assemblies where everyone who cared got together to decide what to do, and eventually they decided to take Google offline and transfer control of its databases to Ma. Some people call it the Second Conversion, but the transition to Ma’s world was no where near as abrupt and difficult as bringing the Attention Economy online. Ma’s arrival feels less to me like a conversion, and more like we’ve finally dropped anchor and set foot on dry land after years at sea. This was the sustainable system we fought so hard for in the reorganization that led to the Conversion, and I get to see it in my lifetime.
Blah blah blah, the future is wonderful. Whatever. My light bulb is still out.
I stayed in bed until noon. This was the third month I’ve been working on The Project, and the next month is going to be even harder if we want to complete it before Labor Day. But today’s a day off, so I enjoyed the comfort of my blankets for a few hours later than usual. I eventually got around to checking in on Ma, who helped located my list of warnings, and I searched for the light bulb. Ma displayed the make and model of the light bulb, and as I had suspected and Ma confirmed, there were no spare light bulbs in my apartment. Ma also provided a list of three places where a replacement could be found, all within walking distance from my apartment. Two of them were private residences, one of which was in my building. I could query Ma for more information, but I could already guess that the one in my building is Mrs. Weasel from 2A. Mrs. Weasel is a hoarder, and always deliberately takes more than personal use because she is lonely and she knows people will eventually have to come to her looking for supplies. I go to her sometimes, especially if I don’t hit the market right when it gets reupped I need something out of stock. Though she would never refuse you access to the supplies, Mrs. Weasel definitely will give you a look that says “Won’t you stay for tea and chat?”, and that makes me nervous. If she tried to block access to the supplies, Twitter would be all over that in a second, and she would probably be restricted from exceeding personal use. So I know I could get a light bulb from Weasel. I just don’t want to have to talk to her.
The last place Ma mentioned was the market at the corner. I needed to eat some breakfast anyway, so I decided to take a trip to the store. Ma’s widget had a list of other suggestions and issues that I glanced over— I didn’t even recognize that I was almost out of toothpaste when I brushed my teeth this morning, but Ma made sure to remind me. At the bottom of the widget was Ma’s signature: a scripted, flowing, “Love, Ma” that was at once delicate and serious. Her signature sat on top of a scrolling field of text overlayed with various graphs and metrics, displaying real time queries Ma was receiving that might be relevant to me— again, nothing was urgent. Ma would tell me if it was. Ma includes this signature in all her messages, which I guess is a way to show us how hard she is working all the time. She is a bit dramatic like that.