I'm intrigued by this idea, and plan to test it to build a new product that is a sibling of an existing one, but with a different targeted purpose. I believe it's verifiable enough to try a goal oriented approach but I'm slightly nervous about this all just being a way to get us to burn the next order of magnitude in tokens!
If you closely work on it, set the right limit, it's unlikely to burn more tokens. There's an obvious downside to using it which is burning tokens - but if you design the loop very well, it's less likely. In fact, that's what Loop Engineering is all about.
I'm intrigued by this idea, and plan to test it to build a new product that is a sibling of an existing one, but with a different targeted purpose. I believe it's verifiable enough to try a goal oriented approach but I'm slightly nervous about this all just being a way to get us to burn the next order of magnitude in tokens!
If you closely work on it, set the right limit, it's unlikely to burn more tokens. There's an obvious downside to using it which is burning tokens - but if you design the loop very well, it's less likely. In fact, that's what Loop Engineering is all about.