Has anyone crossed the persistent memory with automations? It seems like a scalable way to keep track of the status of many automation issues. (like having an automation watching memory to identify gaps in your automations)

We do have a crono automation that check the problems we had in automations runs and try to distil possible solutions in the memory

Yes, the Nightly Memory Curator, I am going to steal this automation, this is perfect

Hey Rajiv, maybe similar topic - is there a also a desire for automations to keep their own memory?

I don't really care about teh mechanics of this, but the world is changing. So when that happens, automations should note it and be able to adapt. Similarly, if there is an anomaly,it would be good for the automation to note it and later we see why and then take approprirate action. I like to think of it as automations remembering state/memory so they can learn from it.

I don't really care about teh mechanics of this, but the world is changing.

Yes

So when that happens, automations should note it and be able to adapt. Similarly, if there is an anomaly,it would be good for the automation to note it and later we see why and then take approprirate action.

Right, so some sort of self-healing / self-improvement functionality that's native to an Automation? Overall I think I understand the direction here, just wanted to have a bit more discussion around it. Thanks!

Ok cool. I can give you more detailed scenarios, but hopefully you get what I am saying.