Raw LLM Responses

It’s always interesting when someone tries to reduce the complexity of GPT and public perception into just two categories: those who “get it,” and those who are somehow deluded by the “Kool-Aid.” But even in your own explanation, a third group shows up: people who confidently misunderstand the tech while criticizing others for doing the same. For example, GPT doesn’t browse the internet in real time unless it’s using a specific tool for that purpose. Most of the time, it generates responses based on patterns learned during training. It’s not compiling sources; it’s producing statistically likely continuations. So if we’re going to debate what GPT is or isn’t, it helps to start with technical accuracy. There’s a deeper issue here too: the question of sentience. You say GPT isn’t sentient as if that’s a solved problem. But we don’t currently have a clear, consistent definition of sentience that checks all the boxes. It needs to include all humans, even those who can’t communicate; exclude obvious non-sentient things like trees or bacteria; and still work if we encounter something alien or artificial. In the absence of such a definition, people will continue to simply claim it is or isn’t - without actually stating what it is or isn’t. This is inherently falsifiable, hence the debate - we’re not debating sentience without a definition; we’re debating how people feel. One person feels like ChatGPT is sentient based on talking to it. Another person feels like it isn’t - based on technical information. I’m not convinced that something as subjective as a random person’s feelings are a reasonable judge of sentience. We need a definition, but for now, we can at least address a few of the common technical criticisms, which may explain why I’m still on the fence. Many of the usual arguments against AI sentience fall apart under closer inspection: **Originality**: OP stated it couldn’t make something original. People remix what they’ve learned all the time. We literally ha