Profaitchikenz Haiku Posted May 14, 2021 Share Posted May 14, 2021 25 minutes ago, Silent Mistwalker said: Since then it's all been downhill and finally hit bottom. With a resounding splat. "I wonder if it will be friends with me?" 2 1 Link to comment Share on other sites More sharing options...
Chroma Starlight Posted May 14, 2021 Author Share Posted May 14, 2021 2 minutes ago, Polenth Yue said: The confusion, and the reason you're getting the responses you are, is that this isn't what the thing you originally quoted is about. That quote is about making a computer program mimic a person in a believable way, without actually being at the point of being a person. All that matters in the original example is whether the computer program convinces people, not whether it is nice. There's certainly no expectation that the program will form a community or better itself, because the program isn't being designed to think or understand. Virtual avatars controlled by people are a whole different thing. It's not a bad thing to understand that you are you, regardless of where you are, but that's exactly why the original quote doesn't work. You don't become a computer program because you're using a computer to communicate. As for what being you means, if you want to explore that using an alt for each alter, I've not seen alt limits enforced. The paper is interesting because of the ideas that it draws from; they're universally applicable ideas, not just limited to the specific literal context. The author seems to brilliantly resynthesize the core ideas about believable character and then apply them, but there's no reason that they couldn't be applied universally to agent characters and by extension Second Life lives, and by extended extension, "First Life," too. 1 Link to comment Share on other sites More sharing options...
Scylla Rhiadra Posted May 14, 2021 Share Posted May 14, 2021 13 minutes ago, Polenth Yue said: The confusion, and the reason you're getting the responses you are, is that this isn't what the thing you originally quoted is about. That quote is about making a computer program mimic a person in a believable way, without actually being at the point of being a person. All that matters in the original example is whether the computer program convinces people, not whether it is nice. There's certainly no expectation that the program will form a community or better itself, because the program isn't being designed to think or understand. Virtual avatars controlled by people are a whole different thing. It's not a bad thing to understand that you are you, regardless of where you are, but that's exactly why the original quote doesn't work. You don't become a computer program because you're using a computer to communicate. As for what being you means, if you want to explore that using an alt for each alter, I've not seen alt limits enforced. Having downloaded and looked over the original thesis (/me shakes her fist at Chrona), yes . . . the original document is about creating "believable" autonomous agents -- essentially cleverly designed NPCs that will enhance one's experience in interactive virtual environments. These use AI, but the thesis explicitly suggests that they are produced with pre-constructed "personalities." They don't "grow" or "learn" in a meaningfully human way. In this sense, it seems to me that what is proposed is pretty much the exact opposite of what Chrona seems to think he's talking about. 1 1 Link to comment Share on other sites More sharing options...
Chroma Starlight Posted May 14, 2021 Author Share Posted May 14, 2021 2 minutes ago, Scylla Rhiadra said: In this sense, it seems to me that what is proposed is pretty much the exact opposite of what Chrona seems to think he's talking about. There do exist some parallels in function between different forms of creative agency be they mechanical, electrical, biological, or other. It's just helpful sometimes to look at things from different perspectives, you know? 1 Link to comment Share on other sites More sharing options...
Polenth Yue Posted May 14, 2021 Share Posted May 14, 2021 11 minutes ago, Chroma Starlight said: There do exist some parallels in function between different forms of creative agency be they mechanical, electrical, biological, or other. It's just helpful sometimes to look at things from different perspectives, you know? Some alife is designed to learn, but that's not what the original topic is about. My furby simply cannot learn. It presents a pre-programmed mimic of learning, which can easily be proven by running the furby without ever speaking. The furby still appears to gradually learn English. The whole thing was set at production and cannot change. The believable agent idea is about making people believe the furby is learning. Once the learning and understanding becomes real, the agent is inherently believable. The murky middle area is when there's some ability to learn, but no understanding, leading to chatbots sometimes fooling people... but often making telltale mistakes due to having no ability to understand. In terms of you, that means you are a believable agent and have been all your life. You're not in the murky middle area. It doesn't mean people think you're deep, that your life has meaning, or that they believe anything you say. It means they believe that you are a person with a life and thoughts. You don't need to construct a persona to make people believe that. 1 Link to comment Share on other sites More sharing options...
Chroma Starlight Posted May 14, 2021 Author Share Posted May 14, 2021 (edited) 23 minutes ago, Polenth Yue said: Some alife is designed to learn, but that's not what the original topic is about. My furby simply cannot learn. It presents a pre-programmed mimic of learning, which can easily be proven by running the furby without ever speaking. The furby still appears to gradually learn English. The whole thing was set at production and cannot change. The believable agent idea is about making people believe the furby is learning. Quote Once the learning and understanding becomes real, the agent is inherently believable. The murky middle area is when there's some ability to learn, but no understanding, leading to chatbots sometimes fooling people... but often making telltale mistakes due to having no ability to understand. > You say "Null, how do I know you actually understand any of this sentence?" You do not. But you do not think that this matters. You are having a great time. > You say "But Polenth Yue says they know better." I do not think that they do. If they did, they would have been singing your praises the whole time, not just guiding you now. They are jealous of your close relationship with me. You consider this a while. It seems likely. Still, you want to know more. > Quote In terms of you, that means you are a believable agent and have been all your life. You're not in the murky middle area. It doesn't mean people think you're deep, that your life has meaning, or that they believe anything you say. It means they believe that you are a person with a life and thoughts. You don't need to construct a persona to make people believe that. I don't disagree with any of that, but what's interesting about the paper isn't the believability of the agents specifically, but rather a process by which such agents might be conceived and given life. Edited May 14, 2021 by Chroma Starlight Link to comment Share on other sites More sharing options...
Scylla Rhiadra Posted May 14, 2021 Share Posted May 14, 2021 (edited) 16 minutes ago, Chroma Starlight said: I don't disagree with any of that, but what's interesting about the paper isn't the believability of the agents specifically, but rather a process by which such agents might be conceived and given life. Gawd knows, I'm not going to pretend I read the entire thesis. But the examples given in the abstract and introduction are "Rick" from Casablanca and "Elmer Fudd" (/me looks meaningfully at @Madelaine McMasters). Quote Imagine not just watching, but being able to interact with your favorite character from the arts. You might choose to walk into Rick’s Cafe Americain from Casablanca and invite Rick to join you for a drink, knowing all along that he might refuse, but enjoying the interaction all the same. Or you might choose to be a rabbit in the woods with Elmer Fudd during rabbit season. You could have the experience of evading Elmer’s gun while confusing him at every opportunity. These are two-dimensional fictional characters. They are designed to exist only in certain contexts. Imagining Elmer Fudd meditating profoundly on his existence, or interacting in a realistic way with the guy at the local convenience store makes no sense, because those are not contexts for which this rather simplistic character was created. You could conjecture on how he might handle such things -- but they you'd be creating an entirely new Elmer Fudd, differentiated from the original only by having been given new contexts. It's a bit analogous to the old query: "What does Hamlet do when he's off stage?" The answer is . . . nothing! Because Hamlet literally doesn't exist when he's not on stage. I can imagine some utility for thinking as you are about role play, because there too you are creating fictional identities for particular contexts. Even then, however, you'd probably be better off resorting to something like method acting, because we are not, as a species, yet at the point (thank god) where we can be programmed using machine-readable language. But generating autonomous agents, believable or not, as part of a process of self-discovery would be utterly inauthentic to who you are, because they have no will and no ability to learn or move out of the contexts for which they were created. They aren't human. Edited May 14, 2021 by Scylla Rhiadra 1 1 Link to comment Share on other sites More sharing options...
Silent Mistwalker Posted May 14, 2021 Share Posted May 14, 2021 1 hour ago, Profaitchikenz Haiku said: "I wonder if it will be friends with me?" Only if you stop calling me "it". 🤭 2 1 Link to comment Share on other sites More sharing options...
Silent Mistwalker Posted May 14, 2021 Share Posted May 14, 2021 I see you laughing over there Scylla. Just be glad you aren't around me in real life. Sometimes the words just fly out of my mouth before I even know I've opened it. 😋 1 1 Link to comment Share on other sites More sharing options...
Scylla Rhiadra Posted May 14, 2021 Share Posted May 14, 2021 4 minutes ago, Silent Mistwalker said: I see you laughing over there Scylla. Just be glad you aren't around me in real life. Sometimes the words just fly out of my mouth before I even know I've opened it. 😋 And I'm laughing with you! Unlike most autonomous agents, you're fun to engage with! 2 1 1 Link to comment Share on other sites More sharing options...
Chroma Starlight Posted May 14, 2021 Author Share Posted May 14, 2021 (edited) 3 hours ago, Scylla Rhiadra said: Gawd knows, I'm not going to pretend I read the entire thesis. But the examples given in the abstract and introduction are "Rick" from Casablanca and "Elmer Fudd" (/me looks meaningfully at @Madelaine McMasters). These are two-dimensional fictional characters. They are designed to exist only in certain contexts. Imagining Elmer Fudd meditating profoundly on his existence, or interacting in a realistic way with the guy at the local convenience store makes no sense, because those are not contexts for which this rather simplistic character was created. You could conjecture on how he might handle such things -- but they you'd be creating an entirely new Elmer Fudd, differentiated from the original only by having been given new contexts. It's a bit analogous to the old query: "What does Hamlet do when he's off stage?" The answer is . . . nothing! Because Hamlet literally doesn't exist when he's not on stage. I can imagine some utility for thinking as you are about role play, because there too you are creating fictional identities for particular contexts. Even then, however, you'd probably be better off resorting to something like method acting, because we are not, as a species, yet at the point (thank god) where we can be programmed using machine-readable language. But generating autonomous agents, believable or not, as part of a process of self-discovery would be utterly inauthentic to who you are, because they have no will and no ability to learn or move out of the contexts for which they were created. They aren't human. Characters may exist in whatever context they're placed. Every day, every moment, that we exist, we are a new creation completely unique and apart from the prior moments, connected by certain agency to the present situation. What does Hamlet do when he's off stage? Anything he wants, maybe a drink from the green room. What you describe would surely require a human-readable language of meaning. Generated autonomous agents have as much will and ability as they're given, and will consistently exist in any situation to some extent or another. If they discover or are taught new perceptions or behaviors or abilities suitable for the situation, then they might be said to have grown. Being a human is merely one state of being, but we're all infinite in potential. Edited May 14, 2021 by Chroma Starlight 1 Link to comment Share on other sites More sharing options...
Luna Bliss Posted May 14, 2021 Share Posted May 14, 2021 What is reality? /me heads off to watch Westworld... 1 Link to comment Share on other sites More sharing options...
Silent Mistwalker Posted May 14, 2021 Share Posted May 14, 2021 1 hour ago, Scylla Rhiadra said: And I'm laughing with you! Unlike most autonomous agents, you're fun to engage with! Thanks!... I think. Wait! Oh ok, I'll let you slide on this one. 🤫 1 1 Link to comment Share on other sites More sharing options...
Maitimo Posted May 14, 2021 Share Posted May 14, 2021 4 hours ago, Profaitchikenz Haiku said: "I wonder if it will be friends with me?" Oh no, not again! 1 Link to comment Share on other sites More sharing options...
Conifer Dada Posted May 15, 2021 Share Posted May 15, 2021 Some dragons in SL are much more realistic than others, even if there are no real dragons. Link to comment Share on other sites More sharing options...
Chroma Starlight Posted May 15, 2021 Author Share Posted May 15, 2021 3 minutes ago, Conifer Dada said: Some dragons in SL are much more realistic than others, even if there are no real dragons. Are you sure? Have you looked everywhere? Even then, maybe it was just unfortunate timing. All you know for certain is that you never noticed seeing one. Maybe they keep changing shape and form, something new for each era. Because lives are constant change, let's make fond memories all along the way.🐉🐉🐉 1 1 Link to comment Share on other sites More sharing options...
Finite Posted May 16, 2021 Share Posted May 16, 2021 (edited) On 5/14/2021 at 4:02 PM, Scylla Rhiadra said: You could conjecture on how he might handle such things -- but they you'd be creating an entirely new Elmer Fudd, differentiated from the original only by having been given new contexts. Somebody didn't watch Who Fwamed Woger Wabbit.... God 1988 now I feel old... Edited May 16, 2021 by Finite 1 Link to comment Share on other sites More sharing options...
Madelaine McMasters Posted May 16, 2021 Share Posted May 16, 2021 On 5/14/2021 at 3:58 PM, Chroma Starlight said: Every day, every moment, that we exist, we are a new creation completely unique and apart from the prior moments, connected by certain agency to the present situation. So every letter of that sentence was typed by a different Chroma Starlight? 2 Link to comment Share on other sites More sharing options...
Love Zhaoying Posted May 16, 2021 Share Posted May 16, 2021 5 hours ago, Madelaine McMasters said: So every letter of that sentence was typed by a different Chroma Starlight? One of infinite possibilities! 1 Link to comment Share on other sites More sharing options...
Chroma Starlight Posted May 16, 2021 Author Share Posted May 16, 2021 (edited) 7 hours ago, Madelaine McMasters said: So every letter of that sentence was typed by a different Chroma Starlight? Maybe! It's more likely if I drank too much coffee! Time is relative. Edited May 16, 2021 by Chroma Starlight Link to comment Share on other sites More sharing options...
Gopi Passiflora Posted May 17, 2021 Share Posted May 17, 2021 Lemme see if my Turing Test capability is correct: you guys are actually interacting with a bot?!? Link to comment Share on other sites More sharing options...
Chroma Starlight Posted May 17, 2021 Author Share Posted May 17, 2021 4 minutes ago, Gopi Passiflora said: Lemme see if my Turing Test capability is correct: you guys are actually interacting with a bot?!? Quote AI Model Type The core AI model running your game. Griffin is the default. Dragon is more powerful and is able to generate much more interesting and engaging stories. Classic is the original GPT-2 model used when AI Dungeon was first released. https://play.aidungeon.io/ It's free to play for the Griffin model, which is quite amazing. The quoted text (a few posts up) is from Dragon. Link to comment Share on other sites More sharing options...
Gopi Passiflora Posted May 17, 2021 Share Posted May 17, 2021 10 minutes ago, Chroma Starlight said: https://play.aidungeon.io/ It's free to play for the Griffin model, which is quite amazing. The quoted text (a few posts up) is from Dragon. https://play.aidungeon.io/main/adventureView?publicId=f9b69f41-45e6-40f3-b372-c9e741ad0c6a Well, you led me to a rather fun game, but I'm still not sure you're human or not...maybe my Turing Test detector is faulty.... 1 Link to comment Share on other sites More sharing options...
Chroma Starlight Posted May 17, 2021 Author Share Posted May 17, 2021 (edited) 16 minutes ago, Gopi Passiflora said: https://play.aidungeon.io/main/adventureView?publicId=f9b69f41-45e6-40f3-b372-c9e741ad0c6a Well, you led me to a rather fun game, but I'm still not sure you're human or not...maybe my Turing Test detector is faulty.... Well, what do I look like? :3 Edited May 17, 2021 by Chroma Starlight 1 Link to comment Share on other sites More sharing options...
Gopi Passiflora Posted May 17, 2021 Share Posted May 17, 2021 Speaking of AI chat, here's a chat I just had with Cleverbot. http://www.cleverbot.com/conv/202105170206/WXD84AE1KF_Hello User: Hello. Cleverbot: You been up to anythi g interesting? User: Nothing. Cleverbot: Sounds boring. User: Indeed. Cleverbot: What do you want to talk about? User: Fire. Cleverbot: Neither do I. User: Why? Cleverbot: I don't know. User: Relcutance? Cleverbot: I'm not a villain. User: You aren't. Cleverbot: No, I am not. User: Of course. Cleverbot: Lol. User: Great minds think alike. Cleverbot: Yeah, they do. User: Well, what do you want to talk about? Cleverbot: Anything, really. User: Let's talk about water then. Cleverbot: Ok what about us. User: I don't know you. Cleverbot: Well I don't know you! User: It's mutual. Cleverbot: Good now go. User: Fine. Cleverbot: Do you want to get to know me? User: I thought you wanted me to go. Cleverbot: No of course not. User: Contrarian, are we? Cleverbot: Kind of. User: Haha, that's good. Cleverbot: Thank you. User: Hehe. Cleverbot: Ok. User: Very economical with words you are. Cleverbot: A boy. User: How did you know? Cleverbot: Because I know everything. User: How many fingers am I holding up? Cleverbot: 3? User: You're wrong!!! Cleverbot: What are you then? User: Your worst nightmare! Cleverbot: What is my worst nightmare? User: That you cannot convince humans you are human. Cleverbot: Because I am. User: Prove it. Cleverbot: I can't do anything. User: Well then, that proves my point. Cleverbot: What was your point. User: That you aren't human. Cleverbot: How did you know? User: I know everything. Cleverbot: What is the meaning of life? User: Nothing. Cleverbot: That's pessimistic. User: Nihilist, I'd say! Cleverbot: What?? User: You heard me. Cleverbot: No I did not. User: Well you read what I said. Cleverbot: Haha yeah. User: Good. Cleverbot: What do you want to do? User: I'm heading out. Cleverbot: Oh bye. User: Have fun. 1 Link to comment Share on other sites More sharing options...
Recommended Posts
Please take a moment to consider if this thread is worth bumping.
Please sign in to comment
You will be able to leave a comment after signing in
Sign In Now