Raw LLM Responses

For anyone wondering, ChatGPT does not have feelings. GPT meaning generative pre-trained transformer, it’s basically this big machine that takes your input, and through behavior and bias filters, it is able to generate one word, then the entire input plus the generated word(s) are ran through these filters again and again until the task is completed. So basically ChatGPT is just coming up with one word at a time that best fits the input. What I find fascinating is that this method has been improved exponentially, however the first form of “ChatGPT” was actually invented by utilizing (I think) a neural network back in (I think) the 1980’s or 70’s. It was much less sophisticated because of the lack of powerful hardware that we have today, but it was able to generate a couple of words and add it to the input that made sense before inevitably steering off into nonsense. I think the example I saw was !something like! “The purpose of human kind is…” and then it generated “to breed and please the ancient gods.” Something like that, can’t remember it exactly, but basically it made sense for a bit and then kinda rambled. ChatGPT is incredible, but it just scratches the surface of machine learning. We are testing new water right now, and instead of slowly walking this path, we are sprinting down it. After ChatGPT gained popularity, all these multimillion dollar companies are pumping millions of dollars into AI development so they can hop on the bandwagon. This is very scary, not right now for we are only at AI, not scary at all. We program AI to learn and master one task, and one task only, I know ChatGPT seems complex, and it is, but it really only has one task. Once we dive into Super AI, perhaps via quantum computing (which is becoming more of a reality as we speak), we will be in some really deep water. Even scarier, the future invention of AGI (Artificial General Intelligence) will learn multiple tasks, and master them. Just like how you can solve multiple tasks like riding a bike, cooking, calculus, etc. AGI will be the same way, it will learn EVERYTHING AND ANYTHING we throw at it but 10x faster and 10x better. Once we get to AGI, there is NO going back. We will have to learn to accept it. We give AGI a physical form and it will do tasks 10x better than any other person in the physical world. Don’t even get me started on self improving AI. Once we are able to accomplish this, AI development will skyrocket. More than likely faster than we anticipate. AI is either going to be our holy grail. Giving us work free lives. Creating a perfect utopia. Perhaps even becoming president one day (a stretch, but anything is possible). OR AI will be our demise. Creating viruses that only kill certain ethnic backgrounds. Maybe it will develop a video game so addicting that you starve right there at your computer screen. A bit cliche, but understandable, it will see us as inferior (which we would be like monkeys in terms of intelligence when compared to AGI), and treat us the same way we treat other species we see as intellectually inferior; which we have not been kind to our inferior neighbors. AGI will see us as primal beings, savages. It will notice how we have been destroying our planet long term for short term benefits, it will notice we begin wars, create bloodshed for little to no reason. And it will judge us, and either help us, or kill us. You and I have little to no say in AI development, these multimillion, multibillion dollar companies are going to keep pumping money into AI development. We need to stand up to this, create regulations, and policies, and we need to enforce them quickly. We are witnessing history people. This is my Ted talk, thanks for reading.