It's interesting that your systems seem to work in a way that an AI would build for itself. I wonder (partially based on your mentioning some downfalls) if your systems would work even better on an AI than a human.
Totally! The fact that current AIs models act as context-limited prompt executors is a perfect fit for these ideas, as the systems described above were, to a large extent, prompt-engineering-based systems. Also, the self-maintenance aspects of these systems are very related to AI coding (self-maintaining codebases), but also could be relevant to AI systems that act somewhat autonomously but have a core of code/writing that they maintain. Hope to write about all of this soon!
This is such a fascinating idea as you've explored it. I love that you mentioned the Memento origin in the original post. I also had my version of this ("root docs" with "roles" that I reflected on daily, weekly, monthly, quarterly)—but it had its own downfall. I'm curious if you've preserved a lightweight version of this system in any form?
Certain parts of the system have persisted: the daily journaling, the seasonal reviews. The aspect to do with being principled has not really been preserved. Though of course I am always trying to figure out what I'm doing, I haven't recently had a central, official "here's how I want to be" in the way that I used to. I'd like to get back to that in some way for sure.
It's interesting that your systems seem to work in a way that an AI would build for itself. I wonder (partially based on your mentioning some downfalls) if your systems would work even better on an AI than a human.
Totally! The fact that current AIs models act as context-limited prompt executors is a perfect fit for these ideas, as the systems described above were, to a large extent, prompt-engineering-based systems. Also, the self-maintenance aspects of these systems are very related to AI coding (self-maintaining codebases), but also could be relevant to AI systems that act somewhat autonomously but have a core of code/writing that they maintain. Hope to write about all of this soon!
This is such a fascinating idea as you've explored it. I love that you mentioned the Memento origin in the original post. I also had my version of this ("root docs" with "roles" that I reflected on daily, weekly, monthly, quarterly)—but it had its own downfall. I'm curious if you've preserved a lightweight version of this system in any form?
Certain parts of the system have persisted: the daily journaling, the seasonal reviews. The aspect to do with being principled has not really been preserved. Though of course I am always trying to figure out what I'm doing, I haven't recently had a central, official "here's how I want to be" in the way that I used to. I'd like to get back to that in some way for sure.