4 ms·
Yeah while these are amusing they really all just amount to people using the tool wrong. Its a language model not an actual AI. stop trying to have meaningful c
by thereddaikon 4y ago
Yeah while these are amusing they really all just amount to people using the tool wrong. Its a language model not an actual AI. stop trying to have meaningful conversations with it. I've had fantastic results just giving it well structured prompts for text. Its great at generating prose.
A fun one is to prompt it to give you the synopsis of a book by an author of your choosing with a few major details. It will spit out several paragraphs and of a coherent plot.
- roywiggins 4y agoMore fun, is to ask it to pretend to be a text adventure based on a book or other well-known work. It will hallucinate a text adventure that's at least thematically relevant, you can type commands like you would a text adventure game, and it will play along. It may not know very much about the work beyond major characters and the approximate setting, but it's remarkably good at it.
- insane_dreamer 4y ago> Yeah while these are amusing they really all just amount to people using the tool wrong. which, when writing any kind of software tool, is exactly what you should assume users will do
- thereddaikon 4y agoYes but due to the nature of ML trained tools that isn't as straightforward as it would otherwise be. OpenAI has gone to great lengths to try and fence in ChatGPT from undesirable response but you can never do this completely. Some of this testing is just that, people probing for the limitations of the model. But a lot of it does seem like people are misusing the software and aren't aware of it. A contributing factor may be how ChatGPT has been portrayed in media in regards to its purpose and capability. As well as people ascribing human-like thinking and emotions to the responses. This is a Chinese room situation.