4 ms·
Not really, though. It's a stranger situation. ChatGPT's language model doesn't remember chatting with you. It doesn't remember anything not in the training se
by skybrian 4y ago
Not really, though. It's a stranger situation.
ChatGPT's language model doesn't remember chatting with you. It doesn't remember anything not in the training set. It's stateless.
It only seems to remember you because the chat transcript gets fed back in again. No "game" is running, most of the time. There's no other "game state."
There is a writer. It can continue stories with convincing fictional characters. But the fictional characters consist entirely of words. This is the same sort of existence as the fictional characters in a novel.
The helpful AI assistant itself is a fictional character created by the writer. Each time you start a chat session, you create another clone.
If you wanted to keep a story going, there are other writers. The writer you're using could be replaced. It should make upgrades easy as technology improves.
Looks like we are going to end up chatting with increasingly intelligent fictional characters who don't actually exist. What sort of meme-creatures consist entirely of words?
- YeGoblynQueenne 4y agoNo hay banda. https://youtu.be/7PlJP56WAhU https://youtu.be/7PlJP56WAhU No hay banda. There is no band. Il n'est pas de orchestra. This is all a tape recording. No hay banda and yet, we hear them. If we want to hear a clarinet. Listen... Un trombon à coulisse. Un trombon con sordina. Sient le son del trombon in sourdine. Sient le son! A muted trumpet. It's all recorded. No hay banda! It's all a tape. Il n'est pas de orchestra. It is an illusion. Listen! https://youtu.be/P1-_k3qjhw0 https://youtu.be/P1-_k3qjhw0
- thequadehunter 4y agoHis argument was that it might not be completely sentient, but could have some properties of sentience that we should be testing for. I'd say the ability to reason well given a unique irl situation is one of them. I understand that it's just predicting text and the model is only working when it receives input, but what if someone took the LLM and made some kind of cognitive loop in the background where the same LLM reads information, thinks to itself when you are gone, and stores these things in its chat content then tells you about the cool stuff it read about the next day. It may be just text, but now it's "thinking". At what point should we start to consider it as a creature with rights?
- skybrian 4y agoAny text could be used a writing prompt. What makes this prompt special? Maybe it's a text in combination with a writer. But you can also swap out writers, so there's a ship of Theseus issue. This ends up being much like the argument about whether a virus is alive.
- thequadehunter 4y agoWhat would it take for you to consider it alive? Or, to simplify...why is a dog "alive" but a chatbot with a cognition loop is not?