Science & Technology Q&A for Kids & Others: Stephen Wolfram Livestreams

Science & Technology Q&A for Kids & Others:

Science & Technology Q&A for Kids (and Others) [Episode 117] »

1 hour 39 minutes

✕ close

Can you talk a bit about Pangaea? How do continents shift? Is it possible to reform Pangaea? Is there technology that could prevent this? How does ChatGPT work? Could you say that Chat GPT has made a graph of the space of words or ideas? I'd love to see a thesaurus based on vectors into semantic space, so you could ask it to give you a word with a meaning close to "A" but heading in the direction of "B." How well does ChatGPT handle slang or figures of speech? Does it understand text as literal, or is it capable of picking up these notes? Could it be that ChatGPT isn't accurate because its training data is text, which may or may not correspond to the real world? Shouldn't we use only real-world data, such as sensory information of sight, sound, touch, taste, smell? With all the new "AI" tools rolling out, what do you think will be the effect on "truth" and "facts" as we know them? Does ChatGPT's ability to mimic emotions means that it is able to feel anything, and how much consciousness does it have? Do you believe physics boundaries need to be coded explicitly, or do you think enough data will result in the model learning principles? How does a neural network experience time? How do all these threads of computation combine to form a whole from its parts? View Less »

Science & Technology Q&A for Kids & Others:

Science & Technology Q&A for Kids (and Others) [Episode 116] »

1 hour 30 minutes

✕ close

Is it possible to produce large amounts of crude oil artificially by manipulating chemical kinetics? How is the distance to a distant galaxy determined? What industries do you think will be most disrupted by ChatGPT and Midjourney AI applications? Thoughts on integrating the Wolfram interface with ChatGPT? An automatic nonsense/falsity detector would be an interesting tool to have. With generative images, Dall-E needs to be able to recycle its output image for incremental improvement. ChatGPT will edit your emails into publishable books. What are some of the most interesting ChatGPT prompts you've come up with that can aid in everyday life? I've used ChatGPT to help give me ideas for movie scripts. Couldn't the falsity detector be used as a way to actually make the output of ChatGPT coherent? Isn't the detector just the necessary component for ChatGPT to learn from its mistakes? How could a language model be integrated with a symbolic system? Do you think ChatGPT is overhyped? Chomsky chuckled about the literally 60 years he's heard we're "on the verge of an AI revolution." Instead of training on text, wouldn't training on the senses that we use, such as video and audio, be better? I suspect that a model that can predict the next video frame will be as intelligent as a human. Video contains text inside itself as well as other degrees of freedom that humans have access to. Do you log all of your keystrokes etc. with the expectation that you will provide this information to an AI to try to understand your thought patterns? I've used it to construct a 365-day nutrition plan; I'm just having my first breakfast based on it! View Less »

Science & Technology Q&A for Kids & Others:

Science & Technology Q&A for Kids (and Others) [Episode 115] »

43 minutes 41 seconds

✕ close

What causes snow? Why doesn't rain just turn into ice? OK, Stephen really knows his stuff on this branch of physics... he's studied in detail. Why is the density of solid water lower than liquid water? Is there any other molecule that also expands when it is solid? Why can you ice skate or ski? What happens at the surface of the ice? For ice skating, the ice melts at the contact of your blade so there is a small layer of liquid water due to friction, and hence it skates. But you also can push and speed up on ice while skating, so an increase in the friction must have happened? Apparently, gecko feet exhibit Casimir-like effects. Stephen, if I may make a suggestion: you need a blackboard (or whiteboard) behind you for the explanations. Is air a molecule? What is air? Can you measure the absorption of CO₂ by plants with this device? View Less »

Science & Technology Q&A for Kids & Others:

Science & Technology Q&A for Kids (and Others) [Episode 114] »

1 hour 48 minutes

✕ close

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? View Less »

Science & Technology Q&A for Kids & Others:

Science & Technology Q&A for Kids (and Others) [Episode 113] »

1 hour 8 minutes

✕ close

Is color discrete or continuous? Why does purple look like violet when it is made of red and blue, which are on opposite ends of the spectrum? Would the spectrum "repeat" itself if we could "see" more of the spectrum? Apparently, there are male folk that can have four visual receptors as well—it's just rare. Can you talk a bit about astronomy? I just got an eight-inch Dobsonian telescope. What should I look at first? Any tips? It's my first telescope, so I'm very excited. Is it possible to discover something unique? Colors are continuous; their perception is discrete. Similar to adding novelty to car lights to cause high contrast, we focus on points; looking there hurts us less. A dark-bright mess hurts peripherals. What causes the northern lights? View Less »

Science & Technology Q&A for Kids & Others:

Science & Technology Q&A for Kids (and Others) [Episode 112] »

1 hour 24 minutes

✕ close

Are there nuclear reactions going on inside our bodies? Do you think we'll ever be able to replace damaged brain parts with computational parts as another form of prosthesis? What ethical implications will become relevant when we combine machine learning and brain sensors/effectors? Suppose a rule creates a memory in our brain. Then it could be an irreducible problem to make a true brain interface for any individual that could interpret a memory or preexisting concept. Truly a fascinating subject. Assuming we are able to completely understand the human brain, one could probably make a complete copy—basically, we could "fork" one brain into multiple copies! Do you think neurons do their signal processing based mostly on discrete states or the temporal difference between states? Even though all brains are different, don't they all "implement" the same underlying ideas? Doesn't this point to some Platonic realm of reality? One of the issues with being able to read and decode a memory is that someone will have the ability to write artificial memories into a brain. It's somewhat scary to think that could happen one day, but it could also be used for good. What about a Turing test, but for memories—like in Inception? Perhaps the only difference between dreams and reality is just a matter of degree? Perhaps it just depends on its logical coherence? Once the logical coherence is larger than what the brain can be aware of, it is considered "real." We've co-evolved with our environment so it should be coherent to us, but if we inject things into our environment that we haven't co-evolved with or evolved in, we get confused. View Less »

Stephen Wolfram Livestreams

Science & Technology Q&A for Kids & Others (153 videos)