[Smarter #2] Write a Book With AI in 43 Seconds, Conquer a Root Canal Relaxed, and Other Smarts
Friends,
You're receiving this Smarter newsletter because you're curious, kind, and changing the world. Plus, I like you.
This newsletter is about all things Smarter: how to save time, make money, and optimize your life. You're a C-suite leader, serial entrepreneur, or world impactor. You know success is about working smarter, not harder.
Who am I again? A healthcare strategist turned tech exec turned consultant. I combine my background, peppered with fancy titles (e.g., Chief Strategy Officer) and notable employers (e.g., Google), to make money for clients. For a fee. Meanwhile, finger-typing 100% human-written Smarter work and life ideas for you. For free.
Just catching up? Start with [Smarter #1]. Just like a good Netflix series (or a urine test sample) catching it mid-stream just makes things harder.
AI Making Me Money While I Sleep
This was Week 1 of my use-AI-to-make-me-passive-income experiment.
First, ChatGPT, Gemini, Perplexity, Claude, DeepSeek, and Manus all confirmed AI would be great at generating passive income. One could argue the sources are biased.
My dream is to get AI to agentically:
a) figure out what people want,
b) identify what Rachel Loui can uniquely offer,
c) make it,
d) sell it, and
e) deposit the income into my bank account.
All while I sleep.
Here’s the reality. I gave AI that exact agentic request. I even tested a handful of agentic models (Manus won) to get the highest-bidding promise. Manus did deliver a product in PDF form. How did it look? I have no idea. The PDF wouldn’t open.
So, I pivoted to a baby-steps approach. Initially, I had the venture-capital mindset: invest in many ideas and hope one becomes the big winner. Why not? AI was supposed to be doing the grunt work, not me.
But then, I made my first AI-generated product (more below). It turned out to be a lot more time-intensive than expected. Imagine an entrepreneur dreaming of opening a bakery to print money like hotcakes. Only to realize revenue comes after leasing a storefront, hiring employees, and baking goods. Then getting stuck in the minutiae of figuring out where to advertise a cashier job posting.
Thus, 98% of this week’s “passive income experiment” was me following (sometimes inaccurate) directions to: format, create seller accounts, upload, and finalize product profiles on online stores.
In other words, AI passive income is possible but not necessarily automatic.
AI vs. Rachel
To amp up the ROI, I’m exploring which products are best as AI + Rachel creations versus those AI can generate solo. The former is arguably more profitable. The latter is more scalable. Even better, the exercise itself could become a future consulting product. After all, which client wouldn’t want to know which AI + Human products protect their jobs, while also seeing which AI-only products could speed them toward retirement?
I prompted AI:
"You are my AI business experiment assistant. I want to run a beginner-friendly A/B experiment comparing two ways of building passive income:
AI-Only (Autonomous): businesses AI creates and runs with almost no human effort (“set and forget”).
AI + Human (Augmented): the same businesses, but with my background, experiences, and oversight. Use my LinkedIn profile (attached as PDF) and anything you can find online about Rachel Loui."
A top suggestion was to create a business ebook. AI promised it would be profitable. I disagree. As someone who published a China-business-bible book in 2016, I can assure you the nine years of Amazon sales profit barely covered the cost of the laptop the book was written on. But hey, who am I to argue with AI about what humans want?
Still, an ebook was the easiest MVP entry point to test:
a) whether AI could create a sellable product autonomously, and
b) whether it could do so comparably (or better) than a human.
After all, half the work of the A/B test was already done. Though, when it comes to popular business topics, it's been almost a decade since China was the 'it' geography. Now everyone is trying to create things not in China. I recognize the experiment's limitation.
I uploaded my Rachel-written 2016 China ebook and told ChatGPT to write a comparable one. The output was, shall we say, not good. The sentences were riddled with ChatGPT’s favorite tell-tale sign—the em dash. And the content didn’t make sense. One line read:
“Tip #3: Eat strategically. Some dishes are symbolism-heavy. Spitting out a fish head? Bad idea. Smile, nod, and remember: culinary diplomacy counts.”
First, why would you put a fish skull in your mouth at a work dinner? And to state the obvious: neither a Chinese nor any other homo sapien stomach can digest a bone the size of a Chinese banquet fish. Of course you’d spit it out. Culinary diplomacy, my arse.
That said, some components did emulate my book pretty well. For example, the double-entendre pen name of my human-written book, Funky Chicken. The pen name was a nod to the funky (amusing) experiences uncovered in a geographically chicken-shaped country called China. It also acknowledged a smart person feeling too chicken to pen their real name while residing in a Communist country that repeatedly wins global jailed-journalist lists.
ChatGPT gave its rendition the author pen name StrategiBot 3000. Just to make it abundantly clear to readers that the content was AI-generated. In case the AI-disclaimer in the product description, copyright page, About the Author page, and profile description weren’t obvious enough.
In total, the time from AI prompt to product sales was five hours. That’s significantly shorter than the full year it took me to write, edit, design a cover for, and publish the human-written version.
For the A/B testing, AI recommended selling the AI and human-generated books on Gumroad for $19.99 (You can purchase here and here at 50% off with the code SMARTER). Which version will win more sales? I have a calendar reminder to check the revenue tally on day 90. You'll be updated then.
In the meantime, next week I'll experiment with AI scraping LinkedIn and Reddit for current unmet needs. Then, creating relevant products people actually want to buy today. What will those look like? Stay tuned to find out.
Other Smarts
Smarter Writing: When starting a newsletter with long content, end with short Smarter tips. Long-followed-by-long content loses readers. Keep. It. Varied.
Smarter Dentist Appointments (And Other Scary Things): When in high-stakes situations, I counter anxiety by relaxing my body. This works anywhere from negotiations to getting a root canal. Imagine when you’re scared. Your body tenses up. Conversely, you can trick your brain into calming down by turning your body to jello. When scared, relax. Really.
Smarter AI-Generated LinkedIn Posts: People ask me how to write AI-generated LinkedIn posts so they don’t look obviously computer-generated. Here are two easy wins. One: em dashes are a mostly no (to my chagrin). Two, emojis are a hard no. Unless you actually wrote with emojis before ChatGPT came out. (And if you were born before 2000, I'm assuming you didn't.) What's the risk of emoji-filling your LI posts? Just ask Netflix's Adolescence. Or your mom when she explores the veggie emojis to write, "Look at that delicious eggplant!" [If you're said mom, here is a cheat sheet]. Instead, try telling AI simply: no em dashes or emojis. Or, keep it low-tech. Just use the delete button.
Until next time,
Rachel
Newsletter forwarded by a friend? Subscribe here.
Rachel Loui is a former Google executive and life sciences Chief Strategy Officer. Today, she uses those experiences to help companies strategically deploy AI and human neuroscience to save time and make money at Strategic Growth Factor. She’s also an improv comedian, international keynote speaker, and book author. Her true passion is trying life experiments, writing about them, and (hopefully) inspiring her readers. Thus, while AI is amazing, this newsletter will always be 100% finger-typed by Rachel Loui. When it comes to human stories, your eyeballs deserve the best.
