When I’m giving it instructions to write code it’s writing down arbitrary things like maybe I told it that I liked a certain function written a certain way but myself You know I usually have code-generated and then move on to the next code… So I don’t really need you know old code memories in there what I was thinking for the memory is it would be better to have it save consistent instructions like if I were to tell it when you’re outputting code for me only output the section that you change and don’t output the entire operator you know like like overall general memories that would improve the operation generally for everything I don’t know if you applied the memory system for people who do code generation and error fixing of code but if my idea would conflict maybe you can make it a tickable option…
I forgot to say what I would do is I would make this a A tickable option and that would be to only save memories if it is specifically requested to update your memory so I would say ‘update your memory In regards to I would rather you only output the specific sections of code that you change instead of the entire operator because I’m a skilled programmer and that’s all the data that I need.’ See that I’m in control and then the quality of your product would increase exponentially for people that know how to give it the proper instruction so if you make that a tickable option so people in the know as far as this could take that and otherwise the functionality have set up could fill in the cracks for people that don’t want to be involved in the thought process. Are you having AGPT assistant analyze the saved memory based upon analyzation of entire sessions and seeing if the memory helped and then automatically optimizing what’s in memory that would be a good idea…