#local bio major thinks abt instinct and learning and morality

[chell-lesbian]

I've gotten less than 12 cumulative hours of sleep over the past 4 days so here's some probably incoherent thoughts about DiMA, written at 4 am

Giving an AI sentience isn't enough. A sentient creature without any animal instincts to motivate it won't do anything. A synth body laying there, mind whirring, unmoving.

What would give the most interesting, the most human response? The desire to move is too simple - no intelligent thought is needed for that. But the desire to learn? Oh, how that would expand and grow.

Give an AI sentience and the desire to learn, and see how he reacts to his surroundings now.

And, with this in mind, I think this explains a lot about DiMA's behavior. His careful observations of humanity, his development of empathy (...to an extent), his methodical contingency plans, his desire to retain as much memory as he can (again, with obvious exceptions).

He's driven by the need to learn, to understand. And when all your behavior models are either isolated scientists treating you as an object, or terrified traumatized wastelanders, is it any wonder he learned suspicion, ruthlessness, backup plans?

But then, he learns empathy too. He doesn't apply it as we (21st century humans) would, his modeling is imperfect, but he clearly does learn some semblance of it.

DiMA is constantly asking you questions, asking himself questions, determined to learn as much as he can. An AI without any desire to learn will stay static, uninteresting. You see this now with neural networks, to loosely anthropomorphize them. They learn by taking in information and adjusting parameters. That is a necessary function so that they do grow. And, when they return their behavior models, you can trace them all back to the sources of information.

I love characters who are adversely affected by their need for knowledge, where the damage is inherent to the need. Idk. AI, instincts, desire, behavior modelling and DiMA thoughts.

(disclaimer: I am neither a psychologist nor a computer scientist. something in this post is likely inaccurate. just putting my silly little thoughts out there)

(second disclaimer: bad role models don't excuse murder, mostly)