Short excerpt from a conversational YouTube video (Tim Davies) discussing the Rabbit AI device. Speakers describe having tried the Rabbit when it launched, note it offered action-capable features via a "rabbit hole" programmable workflow, but that its capabilities were limited and were effectively implemented using a simulated browser (puppeteer).
Got the Rabbit little AI device when it first came out.
Oh, the Rabbit. Yeah, yeah, yeah.
That was a waste of time and that was.
It's quite hardcore techie because I didn't go to that level. And you need to kind of know what you're doing like you can. What could you do with it? I vaguely remember.
Well, not really that much you couldn't already do with the phone, as it turned out.
Oh right.
It was the large action model that was behind it that was appealing. The ability for it to do things as opposed to just come back with words or pict.
Right.
But it hasn't been that well implemented. They built this thing called the Rabbit. I think called it the rabbit hole where you could program in how to take action. So instead of you using apps, you kind of showed it how in a browser to do something. It then remembered that and could.
Isn't that just make. Or if this then that.
Well, it was basically puppeteer. So I think it was just using a simulated web browser behind the scenes to carry out the actions.