My plan was to develop my new creation at work, but not show her to anyone. To make that possible, I needed to get funding. I showed Terry a basic plan which involved taking our currently available online models and upgrading them with the new skin, an automated movement system and a basic AI which controlled both. I also submitted some mocked up focus group responses to the arm, to support the plan. In actual fact I hadn't shown anyone the arm, and had no intention of doing so. After witnessing Terry's completely intense response to it, I didn't want to expose the arm to anyone else. That arm, and the person it would eventually be a part of, was not for commercial release. No one would see her until she was finished.
But I would work on her at the same time I worked on the commercial model. And working on both projects side by side meant I could start delegating responsibilities. I sent out requests to various divisions within the company with attached designs. When the various items came back I would duplicate and re-work them for my own model. The skin I controlled myself, and once I had everything I would handle all the assembly.
This meant that, while I was waiting for various designs to be returned, I had time to work on the software. I decided first to finish the AI for the skin. I had only done rudimentary work on the arm up to now, with controls that would make sense to the arm. A whole body, though, would need more work.
I researched even deeper on skin response - what triggers it and how the responses alter over different parts of the body. I made decisions on every single possible aspect that I could. Not only does the skin on every part of the body respond differently to different situations, each individual also responds differently, and so I had to decide how I wanted her to respond.
Once I was committed to actually doing this - once it became a reality - I realised she needed a name. I had decided to call her Abby. Abigail Whitfield. She was named after two girls I went to school with. One was my first crush in primary school, and the other was a girl from high school who probably never even knew my name, but who all the boys lusted after.
And with a name must come a personality. I spoke to our gaming division about personality simulators and artificial intelligence. They had an AI engine in the pipeline that would be a perfect place to start. There's quite a large adult gaming subculture which gives players the opportunity to simulate seduction and put themselves in the shoes of a ‘babe magnet’. These games started life as very simple interactions, with rudimentary decision trees that, once deciphered, were easily beaten, but in the past few years amazing advances have been made to give these games more and more replayability.
I told the head of gaming what I was doing and he was quite excited by it. He offered to help out with the project but said they were very busy and wouldn't be able to start any time soon. As this was a crucial part of my design, it was one area I wasn't keen to delegate to another department, so I breathed a sigh of relief and asked for any documentation he could give me, which he was more than happy to hand over.
I looked over the documents late one night and began to wonder, what would she be like? What do I want her to be? Who was Abby?
And now, looking over that documentation again, I will have to think what should he be like? Only this time I wouldn't be thinking ‘What do I want?’ because this is not what I want. It's very far from what I want. Instead it's what I must do. I need to do this to get it right. Believe me I wouldn't be doing this if it had all gone well.
I could go into the differences between a man and a woman, and they can be both flippant and serious (and each has as much merit as the other). A friend once told me that the difference was shown in how much luggage was packed for an overnight trip. And we all laughed. But really, those differences - the psychological and sociological - don't actually come into play as much as you might expect when designing an individual. Gender based traits are painted with a very broad brush. They are macro, but individuals are micro. And while when analysing and assessing a person it can be useful to consider the macro, when creating a personality from nothing it can hinder progress dramatically.
Early on, when configuring certain parameters, I found myself asking ‘What would a woman do in this situation?’ and couldn't work out why I was having so much trouble. It wasn't until I started asking ‘What would Abby do?’ that I really started to make progress. Once I thought of her as an individual, requiring individual needs and responses the development of her personality really took off. Of course once I saw her as an individual there was no hope for me. Any chance I had of not falling in love was gone.
And now, when working on this new project, questions like ‘How would a man react in this situation?’ are no more useful. You might think that because I am a man, it's an easier task but it is not. Being a man allows me to ask ‘What would I do?’ and, at best ‘How would men react in this situation?’ But what I should be asking is ‘How would he react?’ Again I need to think on an individual level.
Mass producing these things would be a bitch.
One final aspect of her personality I had to work on was her memories. She couldn't exactly come to life with no memory of anything before. So I gave her memories of growing up, and changing schools. I gave her some pets along the way. That kind of thing. I worked meticulously on creating a background, and also gave her brain the ability to fill in gaps that I hadn't defined.
As I worked on this, the components from the other departments started to come in. I put all work on the commercial model on hold so I could concentrate on Abby. After what seemed like an eternity she was finally coming together.