I asked AI to write the book I never got around to

One of the most intriguing aspects of the current wave of generative AI tools is speed: vast amounts of text can be generated in seconds. Why not write an entire book with the help of AI?

In 2008 I enjoyed a bustling career as a freelance financial writer, addressing topics like retirement investing, wealth management, employer retirement plans, philanthropy and so on. I was not an expert in any of these fields but my clients were, and I picked up enough of the technicalities to act as a sort of translator, interpreting economists’ and portfolio managers’ highly specialized thoughts, for example, to be served on the kitchen tables of ordinary (and extraordinary) investors.

A gap I noticed in the market was exactly this sort of kitchen-table explanation but not on the topic of retirement investing — instead, on retirement income. Sixteen years ago the Baby Boomers had not yet begun to retire and the idea of generating one’s own retirement income was on the radar screens of very few working people. Relevant explanatory material existed, but had not to my knowledge been assembled into an easily digestible book from which the average investor could create his or her own action plan.

I Don't Like Cat Food 2008 book proposal
The original proposal

I wrote a detailed proposal for a book called I Don’t Like Cat Food — an accessible and achievable individual approach to providing for one’s financial needs during retirement.

The central key idea is that investment is like stocking a refrigerator that will feed you during your retirement. Stocking the refrigerator is of course crucial — but so is how you consume what’s in the fridge when you’re no longer working. I was neither expert nor famous, but the proposal seemed to offer a viable approach, I had ready access to experts in the field, and the topic had an enormous and dormant potential market.

Then I spoke with everyone in my network to reach a decision-maker at a mass market business publisher. After months of conversations and meetings, a breakthrough: a business book editor at a major New York publisher was willing to speak with me in a few weeks.

October surprise

The day I was to speak with the editor I had all my ducks in a row: a terrific proposal, an enormous potential market, and a host of client-supporters who backed the concept of my retirement income book project and were willing to be interviewed for a number of subtopics in their areas of expertise.

In October 2008, I phoned the editor as planned. But neither he nor I was in a good mood that morning: the stock markets were in freefall.

Maybe you remember October 15, 2008, when the Dow Jones Industrial Average suffered its then-largest drop, plummeting 733 points in a single day. The NASDAQ plunged almost eight and a half percent, and the S&P 500 collapsed down over nine percent. Obama and McCain were campaigning just weeks before the 2008 November presidential election, and McCain even paused his campaign to assess the financial damage from the market rout. (It would not be clear for a long while what would turn out to be the various factors responsible for the crash — a topic for another time.)

Clearly, the financial world had transformed overnight, and it was anybody’s guess what the way forward might look like. So my conversation with the editor was relatively short and perfunctory. Independent of the events of the day, he didn’t dress up the ultimate truth about the viability of my proposal: they would never publish the business book of somebody without a preexisting platform.

I was not a celebrity and therefore they weren’t interested, no matter how unique and useful the approach was, no matter much potential demand existed in the marketplace, and no matter how superb the text was.

Dispirited because of the rejection, I filed the proposal away and forgot about it.

Enter ChatGPT

Almost sixteen years later we are awash in AI-generated text. Much of it is weird or hallucinatory, but an awful lot of it is quite good! So after quick experiments to see if ChatGPT could write relatively accurate, readable text on financial topics, I learned a simple and personally uncomfortable truth: it could!

I reread my original I Don’t Like Cat Food proposal to see if I felt it was still valid. It was! The full title — I Don’t Like Cat Food: Making Your Money in Retirement Last As Long As You Do: New Strategies for Baby Boomers in a Changing Economy — still needed some work, but the idea, the approach, the logic, the sequence of the material, and the table of contents for the book were solid.

The only reason I never wrote the book was the candor of that editor who said they would never publish a book without some kind of celebrity or expert platform behind it. At the time of the proposal, knowing what I had learned, back then the question was: Why bother?

Today the question is: Why not go ahead and write the book with ChatGPT?

The experiment begins

I will chronicle the experiment as it happens, including prompts, my editing of AI text, creative adjustments along the way, and all other aspects of the writing worth examining.

If it all turns out well, the goal would be to self-publish it on Amazon or by other means. (And if not, no harm done to the readers who will never see it.)


  • AI speed is both an efficiency boost as well as a major editorial risk.
  • Is there a way to harness AI and use it to create worthwhile material instead of insta-generated trash?
  • Writing books is a long and arduous process that tends to benefit celebrity authors who don’t need the money or the additional exposure. Could AI write a book on my idea that would be marketable (or even readable)?

Curious? Contact me. Find me on LinkedIn.

You may find of interest:

Ways to use generative AI text in marketing [2nd in this series]

Beyond perfect [1st in this series]

Q & AI

Claude vs. ChatGPT