Hey guys, just a quick one, I’m trying to set up a system to handle lots of disorganised data for due diligence PE teams, and there’s a lot of it and every answer one may ask about something needs context from 6-10 different scattered files. So my question is using ICM (or not) what is the best way to set up a system/architechture to reliably read and answer from these files? I know ICM can do this just how? The problem is more in the scatteredness of it all

I’m building out an ICM architecture for a PE firm catered towards making due diligence easier and more efficient, I wanted to reach out and see if anyone here is in that space or worked close to it and even any tips you’d have for the correct folder structure/processes. I know a bit but knowing more on your guys’s process would be unreal thanks!

Still tryna wrap my head around this but I want to hear from others why won’t ICM be replaced by better and more efficient AI memory? I’m playing devils advocate here but I really want to learn and get some definitive answers so please discuss

Just wanted to hop on here and thank Jake for reposting my post on using ICM on Linkedin! I'm still learning the real bones of it all but its crazy even with such a simple implementation how much the output quality changes Thank you again Jake for the repost now I'm desperate to learn the next step to document it 🤣 Any suggestions to try to build to further my learning are much appreciated LinkedIn post: https://www.linkedin.com/feed/update/urn:li:activity:7495145654487445505/

Been following this community, as I had been following this similar purpose but without really knowing what it was. Now everyone talks harness, file architecture, second brain la la la. But I still feel we are not there yet. It doesn't feel like this is all, folders and files are cute but how do we look at this from a 5years from now perspective. The dude from claude even says to delete the agents files and let it rip (lol) Thanks