为个人助理造上下文管理系统的开源计划
作者用一年搭上下文管理系统配合 AI agent 管理投资、工作与生活，准备开源，目标是做个人 agent 缺失的那层上下文管理。
spent the last year building systems to run my life because I took on a lot of responsibilities all at once and started juggling too many balls. from 50+ angel investments and managing all of that, to my work at World, my personal life, health, travel, fitness, family, friendships, interests, hobbies, learning, and so much more. I've been very quiet compared to 2 years ago. I am finally getting back on my feet because I've been building a context management system that combined with ai agents is starting to take off a lot of weight off my shoulders. I am in SF for three weeks to lock in and work on world stuff and also shipping the oss version of my context management system. At world I've also mostly switched to building AI systems to scale the foundation and help the TFH (labs) team do the same. If you're building personal or company agentic systems and building/scaling processes hit me up. Really looking forward to launching all the tools I've been building over the last few months. I am working on making them robust and useable for people other than myself right now. the goal is to do for context / life management what openclaw/hermes did for personal agents, i feel like it's the main thing missing from all of these agents and no startup, lab or oss project really does it well. Another topic I'm really into are ontologies, which is the general version of context management. The two best players in this space are Palantir and Databricks. Here are some links to learn more: https://www.palantir.com/platforms/foundry/ - https://www.palantir.com/docs/foundry/ontologies/ontologies-overview/ and https://www.databricks.com/product/platform. The latter ontology project is a more ambitious version of my personal context management system and will take me longer to build, but I hope to launch a really barebones version of it and work on it in the open. I've had tens of conversations with developers at companies I'm friends with about how they design internal agentic systems and this is one of their biggest pain points. We need to have open source ontology engines, together with connectors, visualization libraries and lots of surrounding tooling in order to make agents truly useful at scale, so if you're interested in this as well, hit me up! with the publication of the tools I'm building I'll also publish a long form article on http://dcbuilder.dev/blog about all of the things I've been working, so see you there.