I guess part of the problem is that there are insufficient examples of text-only origami instructions in ChatGPT's set of training texts. We all need to start publishing text-only instructions for lots of traditional and original models to help the AIs of the future be better at this task! Or not. Creating an AI that can create step-by-step text descriptions of what happens on the screen in a video would be another solution, but quite a challenge, I suppose... Best, Matthew Sent from my iPhone
> On 28 Jan 2023, at 17:52, Anna via Origami <origami@lists.digitalorigami.com> > wrote: > > > Laura wrote: > > "Has Chat GPT (I mean, its AI) ever “seen” the process of folding an > > origami crane?" > > Well, I guess it is a philosophical question, whether a text-based AI is able > to see, but I guess no, not in the way we see, feel and experience the world. > As it seems, the AI is still missing data about Origami. > > Louise wrote: > > "Use YouTube videos. The quality varies but keep looking for a video you > can follow." > > Thank you but I'm well aware how to fold a crane. Some years ago Gerwin and I > even "optimized" the folding sequence to be more fun, faster and easier to > fold. You should give it a try: http://www.origami.at/diagrams/crane.pdf > The question was does the newly hyped AI know how to fold a crane too? The > answer is a clear no. AI might be part of our future, but I guess until it > starts to be any useful in Origami it might still take a while. > > Best regards, Anna