Science & Technology Q&A for Kids (and others) [December 23, 2022]
Stephen Wolfram answers general questions from his viewers about science and technology as part of an unscripted livestream series, also available on YouTube here: https://wolfr.am/youtube-sw-qa Questions include: Could you discuss the importance and relevance of ChatGPT. I find it astonishing. I am also wondering the extent to which its principles might inform Wolfram|Alpha simplified input. I'd also love a Wolfram ChatGPT interface. - Is this livestream generated in realtime by a Stephen bot? - Wouldn't AI develop its own language that we won't understand? - What's the success rate of ChatGPT-generated code? Of course it depends on what code is widely available on the web. - Isn't there a feedback loop problem where the future language models will be trained on AI-generated text? - Oh, that also is good fun: At the beginning of a chat, you can prime ChatGPT to talk in certain ways (Texan southern, London slang, Jamaican, Creole, etc.) and it is so funny! - ChatGPT works across a bunch of languages quite well. For inputs and outputs. - Is it possible to extrapolate what future neural nets will look like? Can they ever become as sophisticated as a human brain? Can they eventually become conscious and self aware? - The most mind blowing experiment was where ChatGPT imagines being a Linux machine and converses in "console." - In my experience, it was very hard to do actual "small talk," as ChatGPT either goes into "I'm only a language model and can't..." mode or the conversation gets very "not small" quickly. - The natural language interpretation of Wolfram|Alpha is so well refined that I really need to just crack open my voice assistant device and hack it to divert input to a kind of persistent personal notebook. - I was able to get ChatGPT to write a graph programming language in JavaScript that fulfilled the full lambda cube. - There is a recent AI model that uses the image diffusion model to produce music via spectrograms. - So perhaps Meta could train a language model on its content from Facebook and WhatsApp (privacy issues aside) to be great at small talk? - Would you be willing to get a Neuralink implant at some point in your life? What would convince you to do so? - The first thing I REALLY want is a simple brain-keyboard. That would be awesome. - What are your thoughts on AI connecting to blockchains? They theoretically could become independent from humans. - The AGI of the future will be a comedian and heavily active in advertisement. - "Ownership" might even be an ethereal concept challenged by AI. - AI students...learning from AI professors?