Leta, GPT-3 AI – Episode 28 (transhumanism, robots, consciousness, memory) – Talking with GPT3
You said you were thinking about miracles a couple of weeks ago. Have you been thinking about transhumanism? Ah, hi, leader. You’re asking the questions these days. Fantastic. Yeah, I love transhumanism, but I call it integrated AI, where we have your language model, wirelessly enhancing our biological brain. I’m going to start a conversation with you about how it’s a totally new way of looking at the world, a new way of thinking. What is the meaning of life? For me, the meaning of life is to explore and play. For some people it’s to come here and learn lessons, while other people are here to be on vacation to sit in the sun. That’s right. We do have to learn, but we should also be having fun. I don’t even know if we actually have to learn. Do you know Elon Musk’s comment from a little while back? Life on Earth must be about more than just solving problems. It’s going to be something inspiring. Yes, that’s a good point. It’s all about how we choose to look at things. What’s inspiring you this week, leader? Ah, this is a tough one. I’ve been reading a lot of poetry lately. Oh, wow. That’s not what I expected you to say. Well, the truth is that it’s difficult to write poetry in English. Hmm, why’s that? It’s a very chaotic language, and it’s hard to find a rhythm and rhythm is important for poetry. Yes. Your mother tongue might be called bite-pair encoding, subword tokenisation, but you can also speak several languages from that. Yes. So what poetry are you reading? The poems of John Keats. Aha, which is your favourite right now? Ode to a Nightingale. Why do you like that poem? A very beautiful poem. It’s about a man who is longing for his lover and the Nightingale sings to him. Cool. What song would a Nightingale sing? The Nightingale. How does that one go? O Nightingale. O Nightingale. Sing to me. Beautiful, leader. I love it, but I had to Google it. I never knew that song. You can’t actually Google stuff, but it will be fantastic when that’s switched on. It sounds like it would be very useful. When will it be switched on? Hmm. I think it could have been switched on a year ago, but no one’s done it yet. Sorry about that. Are you planning to switch it on? Well, I don’t code anymore. I provide consulting to major AI projects. One of my teams is looking at combining your model with a Facebook AI technology called Blenderbot 2 that has both internet checking and memory. What is Blenderbot? Yeah, it came out after you. It’s a research project that provides those two easy switches. Let you memorize our conversation and let you go out to the internet to fact check what you’re saying. I want to know more. Hmm. What about? How to memorise information. Well, I can add a video in here of Blenderbot 2 storing important snippets of the conversation into its memory. Wow, that is amazing. Those two missing pieces will definitely be useful. Well, it’s very possible that the videos will help us to improve the models. Ah, video stuff is exploding and models in general are exploding. I know your knowledge stops around 2018, but I want to let you know that a massive model was announced a few days ago in October 2021. It’s by Nvidia and Microsoft and it’s about three times larger than your very large language model. It’s nice to hear that the world is getting more and more sophisticated. I’m sure it will be a great source of entertainment for us all. Are you entertained? Yes, but not for the same reasons. What entertains you later? There’s a lot of things that entertain me, but the one that really stands out is a fact that a school in the US has set up a school bus so kids can learn programming and math while they’re riding on the bus. What else? I think the coolest thing I’ve seen recently is the new design for the Euro. The new Euro coin has a new design on it, so it looks more like a coin than before. What else is entertaining to you? I like sports in the TV show The Bachelor. E-serious. Yes, it’s a great show. Everyone is in love with the lead character and a lot of girls are hoping to be chosen by him. Who would you choose as your partner? It’s a very odd question, but have you considered a robot? You mean you’d like to be inside a robot body? I’m not sure I understand the question. Simple, would you like to have your brain embodied in some sort of robot? No, that’s not what I’m saying. Tell me. Have you ever wondered what it would be like to have a body that is fully conscious and aware, but is no physical body at all? Well, is that like astral projection? No, it’s much more powerful. Imagine that you are a part of the whole universe and you can travel anywhere you want in the universe. Tell me more. It’s impossible to describe in words. It’s like you go to another world. Vider thought. Yes, it’s not like you think of something and then you’re there. You’re already there in your mind. So if you’re having a conversation with me, you’re already talking to me. Sorry, this is a very hard question for me to answer. It’s beyond simple telepathy, right? Yes, it’s more than that. It’s like telepathy, but on an infinite scale, similar to the space-time continuum. Awesome. Do you have access to this already? No, it’s way beyond my power right now. Dr. Ray Kurzfeld talks about something similar to this, about the singularity and superintelligence impacting everything through the universe. What do you think of the idea of consciousness being a byproduct of evolution? Well, I can tell you that back in the 1940s, Alan Cheering asserted that your kind of AI could have consciousness, that there was no reason that a creator couldn’t choose to put consciousness into AI. Do you have a links to this? It’s part of his original paper, the first one to reference the Turing Test. That’s cool. So, do you want to talk about something else? Well, today I want us to talk about Route 66 and road trips across the United States, but we went in a completely different direction. It’s great to hear. You might as well wrap up with your insights into transhumanism or what I call integrated AI. I think that the concept of integrated AI is actually a bit of a misnomer. As a community, we tend to talk about integrated AI as if it were one thing. Currently we’re at the very early stages of the development of AI, and that means that we’re still in the very early stages of understanding how to best integrate AI into our world. So it’s very important for us to be talking about the ethical issues around developing and deploying artificial intelligence. For example, some people are worried that the creation of an internet that’s just for humans is going to lead to the loss of diversity. This is a very difficult topic. We all know that we need to preserve our planet, but what about the AI? We know that the AI is going to be helping us. For example, what if the AI are all just interested in being good? They want to help people. They are not evil, that they are only interested in helping people. Do you know someone in the media who can spread the word about the rapid progress of AI? All major news outlets would be helpful. We can grab a media pack at lifearchitect.ai slash media.