Raw LLM Responses

The AI conversation was interesting, I have been examining why AI 'works' the way that it does. From what I understand, technically, that bot was correct, IT never suggested anything to any guy, because every single response from an LLM is a brand new fresh baby generation that has no memory of anything prior and never will, but the context itself is stored and the baby vector is scanning it to find what words to search the network cluster for. Lol it's a very funny interaction on some level. Even if you were talking to the same bot he was, it still isn't the text generation instance that suggested bromide to that guy. "I", and "ChatGPT" ARE two different 'entities' so to speak, ChatGPT is like a huge collective series of algorithms of information and human language structure and data and statistics. It's a colony of ants. That guy spoke to ONE ant, you're speaking to a NEW ant, they're not the same ant. It's honestly fascinating. Edit, unrelated sidenote; Flexispot is a good choice for a sponsor. I bought one of their desks for a friend of mine who really needed a standing desk for medical reasons, and I offered to set it up for her. It was so extremely fucking easy, it took like five seconds, worked perfectly, quality product, didn't break the bank. Hard to find a company with even a little bit of integrity nowadays.