Use AI to give yourself a second brain

The amount of content I’m consuming (and saving for later) has gotten a bit out of control. It’s hard to keep track of it all. There are tools to save and search things, but I wanted something more customizable, something like a personal library I could actually chat with.

I want to be able to ask questions like: “what was that article about the family and the winery in Spain?” or “what podcasts have I listened to on early-stage investing so I can send them to a friend?”

This week, we’re using AI to build that system. It’s a bit more technical (and a bit longer) of a challenge, but I’ve made it flexible so you can adapt it to your own skill level.

The Challenge:

This week's challenge is different from the previous ones. Instead of asking AI to complete a task, you're going to ask it to architect something: a personal knowledge system that captures everything you save, enriches it automatically, and lets you query it in plain language — like having a conversation with your own reading (or listening) history. Then you are also going to ask AI to help you build it.

The prompt you're running is a design brief. You're the client. AI is the systems architect. What comes back is a custom blueprint built around your habits, your tools, and your tolerance for complexity.

The prompt to get you started

Copy this and run it in your preferred AI system — it will interview you first, then design your system. To open it directly in Claude with the prompt pre-loaded, click here or 👉 Open in ChatGPT

"I want to design a personal knowledge system that helps me capture, process, and actually use high volumes of content I encounter daily. Before designing anything, I need you to interview me. Ask me the following questions one at a time, wait for my answer, then move to the next:

1. What kinds of content do you save most? (articles, newsletters, podcasts, videos, all of the above?) 2. How technical are you willing to get? (no code/automations, light automation like Make.com or Zapier, or comfortable with code and APIs?) 3. How do you want to query your knowledge base? (keyword search, conversational questions and synthesized answers, or both?) 4. What does your reading habit actually look like? (morning digest, throughout the day, or end-of-week batch?)

Once I've answered all four, use my answers to design a custom system. Do not start designing until you have all four answers.

CORE GOAL Turn everything I save into: 1) a clean daily digest I will actually read 2) a long-term, searchable knowledge base I can query in plain language

REQUIREMENTS

1) CAPTURE (must be frictionless) I can save content instantly from anywhere. The system should automatically store the title, source, link, and full text or transcript when possible. This step must be extremely low effort or I won't use it.

2) DAILY DIGEST A short, highly scannable digest that includes key insights and takeaways, a TLDR-style brief for each item, and links back to sources. It should prioritize the most important items, compress low-value content, and be readable in under 10 minutes.

3) KNOWLEDGE BASE All saved content stored in a structured system, enriched with summaries, key ideas, tags, and metadata — and searchable.

4) NATURAL LANGUAGE QUERY I want to be able to ask things like: "What have I saved about X?" or "What were the main themes last week?" and get synthesized answers.

CONSTRAINTS Keep it simple and lightweight to use daily. I don't want to over-engineer or maintain a fragile system. Only build custom where it creates real leverage.

WHAT I WANT FROM YOU 1) Propose the best architecture — what should be off-the-shelf vs custom, and how the pieces connect 2) Recommend a concrete stack (specific tools and technologies) 3) Describe the end-to-end workflow step by step: from saving to digest to storage to querying 4) Show a simple MVP I can build quickly with your help

Optimize for simplicity of daily use, reliability, and speed to implement."

What to Expect:

Your output will be a custom system design. AI will return a full architecture recommendation: which tools to use for capture, how enrichment works, where your content lives, and how to query it. Everyone running this exact prompt will get a completely different system.

Someone who saves mostly newsletters with no automation tolerance might land on a Readwise + Notion setup with a Claude query layer. Someone comfortable with no-code might get a full pipeline that auto-enriches every article with AI summaries and pushes a digest to their inbox every morning.

Too Much?

If the full build feels like too much for this week, start with just the capture piece. In general it's hard to build systems without data, so even starting a simple database of things you'd feel comfortable giving an AI tool access to would be a good starting point.

Try this smaller prompt instead. Open it in Claude here or 👉 Open in ChatGPT

"I want to start saving content more consistently but I'm not sure what system will actually stick for me. Ask me a few questions one at a time to understand my current saving habits, what kinds of content I consume most, how often I realistically review things I've saved, and what tools I'm already using. Then recommend one simple, frictionless way to start capturing content that fits how I actually work. Don't suggest anything that requires a lot of setup — I want something I can start using today."

Why this matters?

Beyond building a useful tool, designing this will give you a real sense of what it feels like to use AI for something bigger than tasks or projects – system architecture.  

It's entirely possible you find this challenge more complex than you need for saving and retrieving articles. That's fine. The goal here is less about the specific output and more about the process: outlining a problem to AI, asking it to design a solution, and then collaborating back and forth until you have something that actually works for you. 

People are using this exact approach to build custom products and workflows every day. The starting point is almost always the same, just clearly outlining a problem and asking AI for help designing a better system. 

That’s it for this week.

Hit reply and tell me what you thought of the challenge this week.

I read every response, and your experience helps me shape these challenges. If you get stuck or have questions, reach me at hello@tinyailab.ai