Why do teams keep losing context, and why hasn't any tool fixed it?
Requirements in Confluence. Architecture decisions in someone's head or a six monthbold Notion page. Code in Git. Slack threads nobody searches. And a new developer joining who has to piece all of it together from scratch every single time.
We talk a lot about building smarter with AI, but the actual bottleneck isn't code generation. It's that by the time AI touches anything, half the reasoning behind the system is already gone. It generates against the "what" while the "why" has completely evaporated.
The result: inconsistent decisions across sprints, onboarding that takes weeks instead of days, and AI suggestions that are locally correct but architecturally wrong.
Has anyone actually solved context continuity on a real engineering team? Curious what's worked or what's failed spectacularly.
3 comments
Depending on your definition of context, my attempt to solve just the foundational context of what words mean in a given domain, and sharing it cleanly, quickly and reliably with both new people in the team and AI agents, is the humble concept map stored in plain text.
Talk here: https://www.infoq.com/presentations/concept-map/
Product here: https://thinkingtools.software/concepticon
Interesting.is this product of yours? Have you used it on a larger engineering team? I'd be interested to know whether it actually reduces onboarding time or improves AI-generated code in practice.
Yes it's a product of mine, but it's actually a recent rebuild in desktop app form of an app I built as a web app well over a decade ago. And yes, I used it to get hmans on the same page at some scale, which you can hear me describe in talk I did at the time like this one:
https://www.infoq.com/presentations/concept-map/
I haven't done formal evals in this AI era about how having a concept map alongside the codebase changes what the agents do, but informally I know mine read them because I tell them to and they seem to take it into account, and I seem to have less trouble than the average geek I talk to keeping my code base adequately coherent. It's just plain text context, after all so there's no magic, just very concise and accurate context.
I haven't done formal evals in this AI era about how having a concept map alongside the codebase changes what the agents do, but informally I know mine read them because I tell them to and they seem to take it into account, and I seem to have less trouble than the average geek I talk to keeping my code base adequately coherent. It's just plain text context, after all so there's no magic, just very concise and accurate context.