Notable economist Tyler Cowen tells us how to use artificial intelligence (AI) for writing and learning. We learn it from an interview with David Perell, in his “How I Write” podcast.
Here is Tyler Cowen’s advice.
- Play around with AI
- Prompting is a skill
- Use all the AI models
- Build your own voice
- AI is the future of the job market
- The increasing value of human networks and human mentors
- Summary
Play around with AI
Cowen advises writers to play around with AI to understand their potential. He uses AI to understand nuances of complex topics. Using AI speeds up his research process: “In the old days,” he said, “I would have ordered and paid for 20 to 30 books. Now, maybe, I’ve ordered two or three books, but I’ll keep interrogating the best LLMs. I acquire the context much more quickly.”
As we all know, AI hallucinates, i.e. it spits out bad info. Cowen prefers to manage the hallucinations. He notices that, over time, the number of hallucinations has decreased. “The very best models with reasoning hallucinate much, much less,” he said. “Hallucinations have gone down by more than 10x in the last year.” To manage it, he advises to verify the information that AI gives you.
Prompting is a skill
A skillful prompt will significantly improve what AI will produce. To obtain high-quality responses from AI, Cowen recommends precise queries rather than general ones.
“Put humans out of your mind,” he said. “Imagine yourself either speaking to an alien or maybe a non-human animal. It’s not a cognitively difficult project to prompt well, but maybe it’s emotionally slightly challenging.”
Use all the AI models
Cowen will switch between several large language models (LLMs), depending on the task. Each model has strengths and weaknesses. Their strengths can be harnessed as tools that suit particular tasks.
For general research, he primarily uses GPT-4 (which he calls “01 Pro”) because it produces reliably deep information with the fewest hallucinations. But when Cowen needs a more creative, philosophical, or nuanced writing style, he turns to Claude. In his experience, Claude produces “a wonderful mix of thoughtful, philosophical, dreamy, flexible, versatile—it’s the best writer.”
For creative writing, Cowen uses DeepSeek, calling it “China boss.” DeepSeek is unpredictable in what it will put out. It will hallucinate more than the other major LLMs, and it will produce weird and emotional thoughts. The playfulness of DeepSeek is a quality that Cowen appreciates reading, but not necessarily using. “DeepSeek is less bland, better at poetry, better at emotion, more romantic, more uneven—it does hallucinate more. It’s just more creative.”
Finally, Cowen often uses Perplexity as an alternative to Google. “Perplexity replaces most of my earlier uses of Google. It’s completely up to date.” Perplexity is an AI-driven search engine. Publishers are getting screwed by the AI search engine bots, however.
Build your own voice
For Cowen, writing in his voice is crucial: “I want the writing to be my own. There are ways you can use AI that will smooth out your writing, but I don’t want to do that. I want to be like Tyler Cowen, this weirdo.”

AI is the future of the job market
The more that people use AI writing, the more important distinctly human stories—those based on authentic, lived experiences—will be. He sees a future where authenticity and personal perspective will be the coin of the realm: “A lot of books I would have written are now obsolete. I think the truly human books will stand out all the more.”
AI is great for student learning. Here’s how Cowen uses AI in the classroom:
“There is no assigned textbook—that saves students money—but they have to subscribe to one of the better AI services, which costs some money, though much less than a textbook would.
The main grade is based on a paper. They’re required to use AI in some fashion and to report what they did, partly because I want to learn from their process. To me, it’s entirely their own work, just like when they use pen and paper or word processing.
I’ve done this before in my law class, requiring at least one paper written using AI. Students recognize there’s a forthcoming equilibrium where using AI will be essential, especially as professionals like lawyers. So why not teach it now? Most students don’t yet fully understand the importance of using the best AI models, and they genuinely want to learn—it’s been a pretty positive experience.”
Academia has lagged in AI adoption. “Every year I ask my econ class, ‘Has anyone else been teaching you how to use AI?’ Silence. And that to me is a scandal. This is academia! We should be at the forefront.”

The increasing value of human networks and human mentors
In the AI robot future, networking and mentorship will be even more valuable. As AI tools become widely accessible, the value of what you, as an individual, knows, will go down. The value of personal connections will increase. Those with strong social networks will use relationships to gain more resources and collaborate effectively.
“Your network of humans is not just like 20% more valuable—it could be 50x more valuable, because the most productive people could be 50x, 5,000x more impactful. They have this free army of highly intelligent servants at their disposal, but to mobilize their projects, they’ll need help from others.”
The future of humans and AI for writing has two pillars: networking and mentorship. Mentors, in Cowen’s view, provide context, practical wisdom, and personalized insights that AI can’t compete with. Mentors help you navigate nuanced and complex decisions:
“Humans know secrets. Maybe AIs can be fed secrets, but they don’t in general know secrets. Now, a human only knows so many secrets—that’s partly where decentralization comes in. How AIs will handle secrets, I think, is a big and interesting question.”
In the end, Cowen has two main pieces of advice:
“Two pieces of general advice, with or without AI in the world, that I think are pretty good for almost everyone, is get more and better mentors, and work every day at improving the quality of your peer network.”
Summary
Practical experimentation with AI tools is vital for success. Be wary of hallucinations and always check the work yourself. Use different LLMs for different purposes. Academia urgently needs to integrate AI literacy into curricula to adequately prepare students for the labor market.
As AI becomes more popular for writers, humans become more important. Cultivating your own voice becomes essential to stand out from the AI crowd. Human-written books will be more highly valued than AI-written ones. Networking and mentorship — with humans — will still be essential to gather information for important life and professional decisions.
Be human in the AI era.

Joshua K. Dubrow is a PhD from The Ohio State University and a Professor of Sociology at the Polish Academy of Sciences.
Also in The Sociology Place
- Artificial Intelligence (AI) and the Social Scientist
- How to use ChatGPT in Social Science Research
- Quantitative Research in the Social Sciences and ChatGPT
- ChatGPT Deep Research Output Misleads Us about Data Collection and Analysis

