ChatGPT Will Be IMPOSSIBLE To Detect (Dangers Of AI) | Lex Fridman Joe Rogan ChatGPT GPT3 AI

Try A Better ChatGPT For Free: 👉 https://bit.ly/BetterChatGPT ChatGPT Will Be IMPOSSIBLE To Detect (Dangers Of AI) | Lex Fridman Joe Rogan #ChatGPT #GPT3 #AI 💥 Make sure to LIKE the video, COMMENT and SUBSCRIBE with notifications ON! 🔔 Watch Every ChatGPT Related Video I Have Made Here: 👉 /playlist/PLjQtCowXP-uTYVJd4xSL5x2JsXIeiKCp6 ChatGPT Is A CODE RED For Google Business (Chat GPT Explained): 👉 /watch/AXf4Tqdktbnk4 This A.I. MANIPULATED Him To Help Her Escape (What Is ChatGPT): 👉 /watch/A_nC3rMNVpPNC In this video, Joe Rogan & Lex Fridman, a scientist and researcher in the fields of artificial intelligence and autonomous vehicles and host of "The Lex Fridman Podcast", will discuss A.I. (Artificial Intelligence), particularly ChatGPT. Lex Fridman explains the basic principles underlying ChatGPT, what large companies are scared of when it comes to using A.I. like ChatGPT and how the potential dangers of A.I. are being handled by people like Sam Altman, CEO of OpenAI, and others, among which is Elon Musk. -------------------------------------------------------------------- Hey, what have your people done, your A.I. people with this Chat GPT. Language models. I don't know if you know what those are, but that's the general systems that underlie ChatGPT and GPT. There's been a lot of development GPT 1, GPT 2, GPT 3, GPT 3.5. in Chat GPT. ChatGPT is based on 175 billion parameter neural network that is GPT 3. Chat GPT 3 came out about two years ago and it was like impressive but dumb in a lot of ways. You can tell it's not intelligent. And what they did with GPT 3.5 is they started adding more and different kinds of datasets there. One of them, probably the smartest neural network currently, is Codex, which is fine tuned for programing. By the way a lot of stuff I'm saying we still don't understand. We're like intuiting why this works so well. Really? These are the intuitions. Yeah, there's a lot of stuff that's not clear. So GPT 3.5 which ChatGPT is likely based on was just trained on code and more data that's able to give it some reasoning then this is really important. It was fine tuned in a supervised way by human labeling. It's basically pointing this giant Titanic of a neural network into the right direction that aligns with the way human beings think and talk. So this supervised labeling allows it to point in the right direction. You have this smart GPT 3.5 thing, generate a bunch of texts and humans label which one seems the best. There's not a significant breakthrough in how much knowledge was learned that was already in GPT-3 and there was much more impressive models already trained. The idea that you don't exactly know why it works the way it works, that that's too close to human, that's too close to human thinking. Like, you know what this is eerily similar to? The plot of Ex Machina. Like, it’s so subtle. That's what ChatGPT is doing. How far away are we from something like Chat GPT being impossible to detect whether or not it's a person or whether it's ChatGPT? Well, it depends who is playing with it. I think we're not that far away in terms of capability, but in order to use these systems and rather in order to train these systems, you have to be a large company. You don't want it to have any kind of controversial opinions. You don't want it to be too edgy. Especially because it's basically an assistant that gives you wisdom about the world, gives you knowledge about the world. If we make the model bigger, 175 billion currently, if you get it to 500, you get it to a trillion parameters, so size of the network grows, size of the dataset grows, is there going to be a point where you're like, holy, what if it starts manipulating you with the answers? It's going to. It's going to manipulate world governments. And that's the cool thing about this is that everyone kind of knows how to do this. It’s computationally difficult, but it's getting cheaper and cheaper and cheaper and cheaper. So, it's not just going to be OpenAI with Microsoft or Google that's doing this, it’s basically anybody can do this. The great thing is the people at the head of openAI currently, Sam Altman and others, really care about this problem. They were there in the beginning. They were the ones like Elon [Musk] screaming about A.I. ethics, A.I. alignment. They're really concerned about superintelligent A.I. taking over. -------------------------------------------------------------------- If you're reading this, write a comment saying "ChatGPT is nerfed" or "GPT-4 will be the best A.I. ever" and I will try my best to find it and give you a heart for it. ❤

الفئة