The AI Brain Dump Dilemma

15/10/2025 31 min

Listen "The AI Brain Dump Dilemma"

Episode Synopsis

Bert shares his battle with AI context windows - that frustrating moment when Claude says "start a new conversation" just as you're making progress. This sparks a deep dive into how different AI tools handle long conversations, when they drift, and practical strategies for staying organized.From creating "cheat sheets" for your AI to organizing projects, backing up your work, and knowing when to start fresh - this episode is packed with actionable advice for anyone using AI regularly. Plus: why AI's memory limitations might actually be a point where humans still win.Bert and Julianna kick off with a Friday recording session. Bert's wearing something other than his signature black polo.Bert introduces the challenge: choosing which AI to use based on context window limitations. Claude cuts you off, ChatGPT slows down, Grok stays focused longer.Where Bert hits the limit: deep architecture and coding work. Going back and forth on "why" decisions, not just "what" to do."It's like I'm calling into tech support... I've spent minutes explaining what's going on and then I have to talk to someone new." The frustration of losing context.Julianna asks: If you're using GPT to organize ideas for a book over three months, does it drift? Bert: "Totally. Drift is a more accurate term."Claude can now check previous conversations. "Let me look" - it searches your chat history to maintain continuity across sessions.Bert connects this to a timeless challenge: big projects are overwhelming. Where do you start? How do you organize context?Bert's tool for huge documents: Google's Notebook LM. Drop in congressional bills, documentation, ask questions, or convert to podcasts.Bert's strategy: Create multiple small conversations about specific aspects. Then make a "cheat sheet" - ask the AI to summarize what you've discussed so you can start fresh.Julianna's observation: Gen Z asks ChatGPT, Millennials Google it. Different approaches to finding information, both creating conversation clutter.Both tools have "Projects" features. Bert's example: Family cartoons for his boys, separate from work. Keeps context siloed with specific instructions per project.Bert's coding workflow: Python script pulls README, file tree, architectural decision records into a single file. Drop it in Claude when starting fresh.Julianna's scenario: Using AI for social media ideas over three months. Setting up projects, giving feedback, going on tangents. How do you stay organized?Bert's recommendation: Save AI summaries to Notion, OneNote, or Google Docs. Ask the AI to create a cheat sheet for itself, then document it.Long conversations slow down progressively. ChatGPT and Grok get slower with more history. Claude cuts you off to maintain speed.Julianna: This is where humans win. A developer or marketer who's been working on something remembers the context. AI is like a new intern every time.If you train AI on the hooks you like, all your hooks sound the same after a year. Humans bring fresh perspectives and challenge themselves to evolve.Bert shares a story of pushing Grok too far. It started forgetting things and led him into a coding corner. Had to go back to manual debugging.Julianna wraps with actionable steps: 1) Clean up miscellaneous chats, 2) Make backups on "digital paper," 3) Attach information to specific projects, 4) Keep living documentation that evolves.