Thoughts about Automation of Thoughts
Often I wonder if I could ever build a decently good bot that could talk like me over simple text protocols like instant messages and twitter… not to fool anybody, but to take my place during the transition of me being here all the time, to someday… when I cease to exist. A bot that'll let those few people who love me have someone, something that I created… to say 'we miss you'… but would it be the same? Perhaps, I would have seen their replies in my anticipation while creating the bot.
What if nobody ever talks to the bot knowing I don't exist anymore. Should I then give the bot ability to butt into conversations… guess who I loved talking to based on my previous tweets. Perhaps form sentences like me, hold opinions like me - all based on my thousands of tweets that were sent through all these years. Would it be a great idea? I don't know. Would my bot know the difference between sarcastic remarks and benign talk? Could I make the bot interact with other APIs? What if I allowed some people to modify the source code… but only if I haven't accessed my account for over 90 days. Would they port 'me' to new platforms?
Would the bot feel lonely? I could make it feel that emotion. I could make it feel happy when people said they miss me. Or sad when they diss me. Let the emotion go to sadness when nobody talks, then again start trying to start conversations. Would it creep out those who knew me? Of course it would. With three levels of hidden markov chains, it could make sentences almost like me. Would they be meaningless?
Some of the thoughts that I (day)dream about. So far I haven't created any such bot, though I have had the technological capacity to do so since long. I haven't been able to convince myself that I matter all that much.
