The AI sparring partner method
Using AI effectively to ask and answer questions isn't just about the answers. It's about thinking better using AI as your partner. Read: "Everything Starts Out Looking Like a Toy" #256
Hi, I’m Greg 👋! I write weekly product essays, including system “handshakes”, the expectations for workflow, and the jobs to be done for data. What is Data Operations? was the first post in the series.
This week’s toy: what did the “everything store” look like in the 1930s? Take a look at the world of the department store, long before the internet.
Edition 256 of this newsletter is here - it’s June 23, 2025.
Thanks for reading! Let me know if there’s a topic you’d like me to cover.
The Big Idea
A short long-form essay about data things
⚙️ The AI sparring partner method
Almost half of UK citizens (47%) have now used a generative AI tool proactively, according to a 2025 Deloitte study. But here's the thing: most people use AI like a glorified search engine. They're asking it to look up information, write emails, or generate basic ideas. And at the same time, studies suggest that chatbot use is rotting our brains and making us less creative.
I think those studies are missing the point.
The problem isn't AI itself. It's how we're using it. When you engage with AI in single-shot queries, you're treating AI like Google. But you do something different when you engage with AI socratically. Building multi-turn conversations engages different parts of your brain in collaborative thinking.
AI might be making me dumber at unaided brain tasks, but I'm also becoming a cyborg with much better powers to solve interesting problems, if I know how to ask the right questions.
We’re using AI incorrectly
The most prevalent AI tasks in this Deloitte study are looking up information, writing emails, or generating ideas. We are treating AI as a tool to get quick answers rather than a partner for collaborative thinking.
The emerging research about AI "rotting our brains" focuses on these basic, transactional interactions. But that's like saying that having conversations with people makes you less creative because you're not thinking alone. The real power comes from the collaborative thinking process itself.
The question is: how do we move beyond this basic usage pattern?
Using AI with the Socratic method
What if we thought about AI not as a tool, but as a sparring partner? We need a partner to challenge assumptions, explore different angles, and push our thinking in new directions.
Let’s call it "socratic AI”. We can learn faster by using AI to engage in the kind of back-and-forth dialogue that Socrates used to help his students discover knowledge for themselves.
Don’t think of it as getting answers from AI. Think of this process as learning how to think more effectively because you’ve had more chances to sharpen your argument with a chatbot.
But how do you actually do this?
Moving from “one-shot” questions to multi-turn
Here are the key techniques for moving from using AI like Google to using it as a sparring partner:
1. Start with raw notes, not polished questions
Instead of trying to craft the perfect prompt, start with your raw thoughts. Create a markdown file with your initial ideas, even if they're just bullet points or rough outlines.
2. Plan before executing
Don't just ask AI to do something—ask it to make a plan first. "Please make a simple plan that shows what you will do and suggest it to me before you execute."
3. Don’t forget about the context!
Tools like Cursor (or projects in ChatGPT/Claude) let you provide a folder of related information as context. This gives AI a richer understanding of your work.
4. Challenge AI's confidence
Learn to adversarially prompt AI to show you the limits of its knowledge. The SIFT toolbox by Mike Caulfield is an excellent resource for this - it uses AI to examine items from multiple viewpoints.
What does this look like in practice?
A real-world example that you can try
Here’s a simple example of how to start using this method. I wanted to manage time better, but didn’t want to use a traditional project management system like Asana.
To begin, I created some raw notes:
Need better daily planning
Don't want complex project management
Want to track progress on bigger projects
Need weekly summaries
Something simple I'll actually use
The first AI suggestion was too complex, so I asked for something simpler and more flexible. Through a few quick exchanges, we landed on a daily log template I could fill out at the end of each day, with a weekly manual summary. The AI helped me personalize the template, adding sections like "Today's Wins" and "Tomorrow's Focus," and suggested ways to build the habit into my routine.
This back-and-forth took about 10 minutes and resulted in a system tailored to my preferences. It was simple, actionable, and mine.
The key insight: I didn't just get a generic solution, but one that fit how I actually work, thanks to the collaborative process.
This is the difference between using AI like Google ("give me a time management system") and using it as a sparring partner ("help me figure out what would work for me").
Teaching this practice through questioning
To teach these techniques, start with non-technical examples and focus on the collaborative thinking process. The most important skill is learning to ask better questions.
Instead of asking directly how to solve and making it your only question, try prompts like "what are the different ways to think about this?" or "what would be the best approach for someone like me?"
Watch this video from Northwestern University's 2025 commencement:
What’s the takeaway? The AI Sparring Partner Method isn't just about using AI better—it's about thinking better with AI as your partner. The real question is: "How do I learn to think with AI as my sparring partner?" The answer starts with learning to ask better questions.
Links for Reading and Sharing
These are links that caught my 👀
1/ The new “everything store” - Brad Stone’s 2013 book The Everything Store chronicled the rise of Amazon.com, a place to purchase anything. The new place to find everything might be ChatGPT or Claude or Gemini, because they are general purpose software. Instead of finding the exact intent, the goal is to build a habit where eventually, you’ll figure out what you’re looking for (or the AI chatbot will find an intent for you that resonates).
2/ is AI chat rotting your brain? - That’s the premise behind a new story from 404 Media about research that demonstrates that people perform worse on unaided tasks than they did before they used a chatbot. One missing perspective here is to drill in and understand whether these test subjects used chatbots like a search engine or a socratic partner. I can totally buy the idea that brains might get lazy after receiving an instant response to any question. But that also begs the question: what if talking to chatbots made you sharper at asking the kind of questions that brought you to the right answer more quickly? That doesn’t sound like brain rot, but more like adaptation to a new way of thinking.
3/ Is there a middle path to the future? - I thoroughly enjoyed (and was also a little bit freaked out) by this conversation from the Center for Humane Technology on the possible futures for AI. It’s worth your time to think about a possible future that’s neither dystopian nor luddite, but models the exponential change that will happen (and is happening already) in the context of social change.
What to do next
Hit reply if you’ve got links to share, data stories, or want to say hello.
The next big thing always starts out being dismissed as a “toy.” - Chris Dixon