The great AI experiment (or how four AIs walked into a bar...)

Hello fellow adventurers!

You know how I mentioned I've been spending far too many late nights tumbling down YouTube rabbit holes? Well, this past weekend delivered what can only be described as a eureka moment wrapped in the glow of my laptop screen at an hour when I should have just gone to bed.

I stumbled across a few videos talking about free one-page websites that solve a single pain point for people. These sites get monetised through Google AdSense, and apparently, some early adopters have made serious money. Like, "buy your husband that really super duper nice birthday present and then some" kind of money.

I was immediately intrigued on multiple fronts. First, that sounds like fantastic passive income if you can nail the concept. Even a few dollars a week beats my current earnings of exactly zero. But it hit me pretty quickly that most obvious ideas are already very, very done. The smart brains have grabbed all the low-hanging fruit and probably the medium-hanging fruit too.

So, the million-dollar questions became: How do I find an idea that might actually work? And, how on earth would I execute it when I can barely spell "HTML"?

But here's the thing, I keep hearing about people with zero coding experience creating hugely successful AI-led businesses. Stay-at-home parents building apps that take off. Regular humans doing extraordinary things with a little AI assistance. Surely there's room for one more (like me!) in that group, right?

Sunday Morning Inspiration Strikes

I woke up Sunday morning with a plan brewing. I'd been toying with an experiment for a while, but now I had the perfect project to test it with. I was going to rally all my AI “buddies”—Claude, Copilot, Gemini, and ChatGPT—and see where their combined powers could take me.

My approach was simple. I crafted this guidance:

"I want to create a simple, free one-page website that I can monetise with Google AdSense. The website needs to: 1) solve a single pain point that people search Google for, 2) be achievable for someone with no coding skills, 3) have no potential legal liability issues, and 4) feature easily in the top five search results, so ideally not an oversaturated idea. Please give me your top five ideas."

Then I hit enter.

And waited for brilliance to unfold…

The Reality Check Moment

It was so much fun! Right up until I realised I didn't like any of the ideas. They all read terribly! I was expecting genius-level innovation, and perhaps there was something hidden in the suggestions that wasn't obvious to my untrained eye, but nothing vibed with me AT ALL.

It was fascinating to see how each tool presented its findings, all very convincing, very confident. But then some suggestions started seeming a bit... done. As in, "I've definitely seen this exact thing before" done.

So, I pushed back: "What are your next five top ideas based on the criteria I gave you?"

I was still underwhelmed by the fact that four AI tools supposedly revolutionising the world weren't exactly wowing me with their brilliance in this particular moment. Time for round three: "Out of the ten ideas you've given me, what is your top pick that best meets my criteria?"

The Battle of the Best Ideas

Each tool gave me their champion and explained their reasoning. Copilot's winning idea? "Sunrise & Sunset Times for [One Specific Location] – [Year]." It sounded interesting, familiar, but I figured there must be a good reason for the confidence.

To take the experiment further, I decided to pit them against each other. I gave each tool the others' top picks and asked them to consider all options before declaring their ultimate choice:

• ChatGPT championed a Laundry Symbols Decoder site

• Claude backed a Sourdough Starter Feeding Calculator

• Gemini went with a Calorie Counter for Fast-Food Items

• Copilot promoting its sunrise/sunset times idea

Then I asked each to play Google specialist and pick the absolute winner from the group.

The Epic Fail Moment

Here's where it got interesting. Copilot revised its position and pivoted to Claude's sourdough idea with no encouragement from me, while the other three stubbornly stuck with their original picks. But then came my genuine LOL moment when Gemini casually dropped this bomb:

"Google itself provides the sunrise and sunset times directly on the search results page (SERP). This is called a 'zero-click' search..."

Wait, what? So Copilot had failed spectacularly on that front! I tested it myself and sure enough—there were my sunrise and sunset times displayed right in Google's search results without needing to visit any website at all.

Thanks for playing, Copilot. Better luck next time!

The Echo Chamber Question

I went back and forth with each tool, asking for clarity on ideas that seemed promising, then doing my own googling to test whether I was being fed charm with little substance. My husband keeps pondering whether these tools sometimes create a large echo chamber where we only encounter information that reflects our own biases, and honestly? Sometimes it does seem that way.

My goal with using all four tools was to remove some bias and hopefully get more insight into their individual strengths, while using them to validate each other's suggestions. (Like Gemini's brilliant Copilot reality check)

The Late-Night Lightbulb

Then I decide on a new strategy, I would pitch my own idea. I'd fallen asleep the night before with a particular idea rattling around my brain, wondering if it would work and fit my brief. After getting nowhere particularly useful with the AI suggestions and not being super keen to launch into the unknown world of sourdough calculations (which the majority seemed convinced was both niche and solid), I threw my late-night thought into the AI wilderness.

And the response was actually really good!

I did my own searching (don't forget to keep verifying, people!) just to see what I thought, and it started looking like this could actually be a thing. So, I asked each AI tool where to start, and the responses were fascinatingly different:

• One gave me a few quick steps then launched into a long block of code that absolutely wowed me (so this is how the "non-coder using code" thing happens!)

• Another focussed on marketing and SEO suggestions

• Claude gave me a comprehensive plan from Day One to Launch Day with a systematic approach to work through

Thanks, Claude, that was super helpful!

The Adventure Begins (Again)

So, my mission now is to see how far I can take this. What do I do first? Honestly, no idea. I need to study Claude's plan and cross-reference it with what the other tools provided. I have absolutely no clue what this will end up looking like, whether Google AdSense will accept me, or if I can actually make money doing this (let alone if I discover it really is beyond my current skill set and abandon the idea altogether!).

But here's what I do know: I'm going to keep learning, keep pushing, keep experimenting, and keep having fun! Because reinvention is sometimes about trying a lot of different things until you find the one that actually works. And, at the end of this journey, I might just be a non-coder who built a coded website!

How wild is that?

Bring on this week, I say. Let's see what happens when curiosity meets code, AI meets ambition, and Dallas 2.0 meets her first substantial technical challenge.

Spoiler alert: I have absolutely no idea what I'm doing, and I couldn't be more excited about it!

Stay tuned for the next instalment of "Will she actually pull this off?"

Dallas 2.0 🚀

Next up: The plan, the execution, and probably some spectacular fails along the way. Because what's a good adventure without a few plot twists?