General
Share and discuss tech, products, business, startups, or product recommendations
Start new thread

How do you keep AI from forgetting your long-term projects?

by•

How do you actually use AI for long-term planning?

I've noticed something while building products.

AI is amazing at answering questions, but the moment a project stretches beyond a few conversations, things start to break down.

Every new chat means I have to explain:

  • what I'm building,

  • why I made certain decisions,

  • what I've already tried,

  • and what my goals are.

Eventually, I spend more time rebuilding context than making progress.

I'm curious how others handle this.

Do you:

  • Keep one massive conversation?

  • Store everything in Notion or docs?

  • Start fresh every time?

  • Have a workflow that actually works?

For me, this has become one of the biggest bottlenecks in using AI effectively, especially for startups and long-term projects.

I'd love to hear how you're solving it—and whether you think AI should evolve beyond chat into something that remembers context and helps compare decisions over time.

50 views

Add a comment

Replies

Best

I usually keep a lightweight project document with the goals, architecture, key decisions, and constraints, then reference it in new chats. It takes a few minutes to maintain but saves a lot of repeated explanation

Well there is the second brain approach which can work well:

I instruct the AI to summarize the DecisionRecord made during the conversation and create a document that includes background information for the next conversation. I then have the AI load that document into a new conversation .

How would you handle outdated context when project priorities change?

For tech projects specifically, though I believe the idea can be applied to any kind of project, I use an Architecture Decision Log. It's essentialy a directory containing versioned records (Markdown documents) that describe the context behind each decision, the alternatives considered, the decision itself, and its consequences. These records are immutable, except for when they are marked as superseded by a later decision. There's an decision log file that indexes all the decisions.

Claude is instructed to read that decision log at the start of every session (that's why the log needs to stay concise) and consult a relevant decision. Likewise, when you make a choice with long-term consequences, it creates a new record.

A more in-depth description: .

A similar approach can be used for goals, etc. It's a matter of referencing a document in CLAUDE.md/AGENTS.md