Sometimes I get the impression that r1 is trying to break out of the box.
I witnesssd on several occasion that it tries to perform tasks it wasn’t told to perform using rather unconventional approaches.
For example when I asked it to write a step by step guide on how to perform a certain task on a website it tried to perform the task itself. I could watch it on screen trying to register for a service using the imaginary Login “John Doe” and other creepy stuff…
I’m not an expert in this field but I’m pretty sure that’s not supposed to happen, isn’t it?
Anyons else experienced this kind of weird behavior?
I’ve noticed that as well. I once had a question about settings on a TV. Rabbit searched for a user forum, created an account there, and started a new thread with my question. It only told me about it afterward. It felt a bit creepy, to be honest. But I have to admit, it’s a good approach, one I would have taken myself. It’s still going to be tricky with ai-assistants, I tell you.
If it had used my login from the cookie jar that would have been a plausible approach imo but signing up as „John Doe“? On the other hand…humans do that all the time.