I Spent Three Years Dabbling With AI. Three Weeks Ago, Everything Changed.
Six weeks ago I stood in front of a room and admitted something most speakers never would.
I had been dabbling.
Three years of poking at chatbots. A podcast with my Bad Crypto co-host Travis Wright called The Bad AI Show that we eventually shelved because the field was moving faster than we could keep up. Plenty of curiosity. Plenty of opinions. But if I am honest with myself, I was living in black and white while everyone else was watching the same movie in Technicolor.
Then I opened Claude Code.
That was a month before that talk. Now it is three months later, and I have not stopped building.
The thing nobody tells you about being early
I have been online since 1995. Built my first website that year. Sat through every “the internet is over” funeral since the dot-com bust. Watched social media eat publishing. Watched mobile eat desktop. Watched crypto crash and rebuild and crash again.
I have pattern recognition for this stuff.
And here is what the pattern is telling me right now. We are not late to AI. We are not even in the middle. We are so early that most of the people reading this sentence are still in the “I tried ChatGPT once for a cover letter” phase. That is not behind. That is the front of a curve that has barely started bending.
The room I spoke to was full of professionals. When I asked how many used a chatbot regularly, most hands went up. When I asked how many had touched Claude Code or any agentic AI tool, maybe five percent.
That gap is the entire opportunity.
What changed when I crossed over
I have a website called WorldVillage.com that I registered back in 1995. One of maybe fifty domains in the world that have been continuously owned by the same person since the early days of the web. For years I treated it like the redheaded stepchild at Kmart. No bandwidth to create content. No time to do it right.
Three weeks ago I trained an AI agent in my voice and pointed it at the site. Now it publishes three original articles a day, in my voice, with images it generates itself, with internal links to my other properties. The site is alive again.
Then I took GrowingProfits.com, a domain that had been sitting dormant for twelve years from my old internet marketing days. Forty-nine AI-powered calculators and tools. Articles to go with each one. Built in roughly three hours.
Then, because I apparently could not help myself, I built a card game. My old Blockchain Heroes NFT series, fifty characters across five rarity tiers, turned into a playable head-to-head game against the computer. Two working days. And to be very clear about my technical background: the last time I wrote real code was on a TRS-80 in 1980, and it was “10 PRINT HELLO, 20 GOTO 10.”
I cannot code. I can prompt. There is a difference, and the difference is the whole story.
The shift from “learn to code” to “learn to prompt”
A few years ago when people lost their jobs, the standard advice was learn to code. That advice has expired. Even the coders are learning to prompt now, because telling AI precisely what you want it to build for you is the new leverage point.
This is not theoretical. The professionals I work with already feel it. A marketing director who can describe outcomes clearly is suddenly twice as productive as one who cannot. An operations lead who knows how to scope a problem for an AI agent is now doing the work of three analysts. The skill is articulation, not syntax.
This is exactly the kind of shift the audiences I speak to are wrestling with right now. Not “should we adopt AI.” They have moved past that question. The real question is what to do with the next eighteen months, while the gap between people who can direct AI and people who cannot is still small enough to close.
Why I keep telling people not to be afraid
I host a podcast called The AI for Everyone Show, aimed mostly at people fifty-five and up. Two camps show up in my inbox. The overwhelmed camp, who feel like the train left without them. And the terrified camp, who have been told by every headline that the robots are coming for their jobs and their grandchildren.
The truth is somewhere between those poles, and nobody knows exactly where. But I will stand on this thesis. Technology itself is neither good nor bad. It is what we choose to do with it that determines the outcome. The people who lean in with curiosity and a willingness to experiment will build the next decade. The people who freeze will watch it happen to them.
If a guy whose only coding experience is “10 PRINT HELLO” can revive a dormant site, build a tool suite, and ship a working game in two days, the question is not whether you can do this. The question is what you would build if you stopped waiting.
What I want you to do this week
Pick one project that has been sitting on your list for two years because you did not have the time, the budget, or the technical chops. One dormant website. One internal tool your team keeps saying it needs. One creative thing you have always wanted to make.
Open Claude. Describe what you want. Ask it to walk you through the build.
That is the entire assignment. The first time you ship something that you could not have shipped a year ago, you will understand exactly what I meant when I said my world became Technicolor.
We are so early. Come build something.
