Theles Spinner

[Varil]

There's a thought for an alternate activity. Build a robot, and have it judged based in things like how it's AI rates, how generally strong it is, and 'aesthetic appeal', which could be beauty in the same way a car might be considered 'beautiful'. Not a straight beauty contest, but...

Maybe a 'kiddy' version, judging robots less on their straight ability, and more on their general appeal to the populace. The Puppy Sim 3000 is generally useless, but durn it's cute, and kids love it for it's realistic interaction!

[palefire]

That's what 'they' want you to think. I bet the Three Laws of Robotics is never built into the "puppies".

It's strange that we are talking about alternative activities in the Theles Spinner thread and Theles Spinner in the alternative activities thread.

I have always wondered why the sentinents robots built in GH1 are so far more charismatic than their creators. On the other hand robotics is a skill taught at the university in the game, yet you don't get to see robots very often in civilized places.

[Varil]

Honestly, I just open all the links with new posts, then read said new posts, and respond accordingly. I don't pay a great deal of attention to the original topic. *shrug*


I imagine robots would either have nil charisma or *great* charisma, depending on whether they are programmed for it. With a perfect memory and blazingly fast logic, not to mention possible senses well beyond mere vision and sound, it seems reasonable that they would be masters of saying what people want to hear, or for leading people down a conversational path to get the information they need.

Or the GH stat calculations might just default to 'uber' for sentient robots. Whatever.

[macksting]

It's not so much default as oddly weighted. Sentient robots can't have a stat lower than 10. By the time you succeed at making a robot with no stat lower than 10, the stats are probably pretty ruttin' high.

As to perfect memory and logic, I can assure you that these do not make a charismatic person by their lonesome. While the truly charismatic are often intelligent enough to know how to make it work for them, they are not necessarily logical at all; some are totally batshit. More importantly, knowing how to do something doesn't necessarily confer the technique of doing so.
For example, while a person who had the knowledge of how to do CPR could attempt it and, perhaps, keep somebody alive until medical assistance arrived, his chances would be lower than somebody who was actually skilled in the doing of such. More importantly, there's a difference between knowing how to land a plane by instruments and successfully doing so.
In the Fuzion system used in the Bubblegum Crisis RPG, the stats which denote this are Intelligence and Technique (INT and TECH), the Int used to design and the Tech used to create.

[Varil]

Maybe when sentient robots are added to GH2, they should be given a point-based system, similar to normal character creation? Auto-distribute some, and let the player decide on the rest, with robotics skill determining what percentage you get to distribute yourself, and skill and material determining how many points are given.


On machine charisma...
If you could read the thoughts of everyone around you, not through telepathy, but through recognition of every subtle sign and inkling of body-language telling you what they feel, do you think you could play to their emotions? Assume you could keep a straight face and use any tone or level of voice you want, and that you can perfectly mimic whatever body posture you need. A sentient humanoid machine would have access to all of this. The conversation skill in game mimics 'technique', but charm is their raw ability to be liked, to use empathy to get where they need to be.

True, a human politician could talk the pants off a machine with no experience interacting with humans beyond base programming and very limited interaction(with it's creator, for instance). However, come back in 10 years after the AI has spent significant time learning about humanity, and it'll out-talk and out-impress any given politician without trying. A machine could read and react to a person far better than another person, giving the ability to compile a sort of psychological profile on-the-fly, which would be a huge edge in a conversation.

It isn't charm, or charisma in the standard sense of being someone people naturally like, but it leads to the same thing and works in much the same way. A person probably couldn't pull it off because we can't mask our emotions very well, usually, and we can't take in every twitch or shift in the other person to determine what they could be thinking.

[palefire]

I was thinking more along another evil line of explanation based on utility for the creator of the robot but my theory is flawed because the robots don't necessary turn out to be the opposite sex of the creator, if you understand what I am talking about.

[Varil]

The Naughty-Bot 4k, with kung-fu grip?

[macksting]

Quoting: Varil

On machine charisma...
If you could read the thoughts of everyone around you, not through telepathy, but through recognition of every subtle sign and inkling of body-language telling you what they feel, do you think you could play to their emotions? Assume you could keep a straight face and use any tone or level of voice you want, and that you can perfectly mimic whatever body posture you need. A sentient humanoid machine would have access to all of this. The conversation skill in game mimics 'technique', but charm is their raw ability to be liked, to use empathy to get where they need to be.

On human perception...
Mirror Neurons. Granted it has the pitfall of imperfect control of empathy, but it's vastly more efficient than slowly plodding through the sequences and dossiers of a changing human landscape.
For that matter, without actual empathy, I'm not sure a machine can even pass a rutting Turing test.

Frankly, I think the method you describe is far too clunky to be put into practice any time soon. Even once something like it becomes possible in realtime, there are simply more effective ways of arriving at a similar result.
In this case, that similar result is roughly that of giving the robot an anima; in realistic terms, I think we could assume GearHead robots to have something similar to mirror neurons.