The Leap of Faith Quandary

I remember walking down Yonge Street in Toronto many years ago, talking on a transportable cellular telephone.

Think about that for a moment.

I was walking down a city street, having a conversation with someone who was somewhere else, without a telephone wire connecting us.

Today, nobody would give that a second thought. But back then? It was science fiction coming to life.

I’ve experienced a lot of technological breakthroughs since then. The internet. Smartphones. Computers that fit in your pocket. Watches that can make phone calls. Instant access to practically all the world’s information.

But recently, something happened that made that first cellular phone experience seem almost ordinary.

I got into a Tesla, activated Full Self-Driving (Supervised), and let the car drive me somewhere.

And I was absolutely blown away by the experience.

It was a magic carpet ride. My childhood imagination come to life.

Here was a machine navigating real roads, responding to traffic, making decisions, and taking me toward my destination. I was still responsible for supervising it and being ready to intervene, of course. This wasn’t a driverless vehicle. (Yes, I know about Cybercab and Waymo, but I haven’t tried them yet.)

Nonetheless, the experience was extraordinary.

And here’s something that surprised even me, Mr. Tech Lover.

The psychological leap of faith required to trust that car wasn’t much different from the leap required to let artificial intelligence operate my computer.

The potential consequences were certainly different. A mistake on my computer might cost me time, money, or data. A mistake on the road could cost me my life.

But psychologically, both experiences raised the same fundamental question.

Am I really prepared to let a machine take control?

There’s a big difference between answering and doing

When ChatGPT first appeared, I was fascinated.

I started experimenting with it almost immediately. I asked questions, explored ideas, wrote content, and discovered that it could be an extraordinary thinking partner.

But there wasn’t much danger involved.

If it gave me a bad answer or a stupid picture, I could ignore it. If it made something up, I could challenge it. It was essentially a remarkably sophisticated conversational tool.

Then something changed.

Artificial intelligence began acquiring the ability to act.

Not just tell me how to do something, but actually do it.

Open files. Operate software. Navigate websites. Change settings. Work with information. Execute tasks.

This was an entirely different proposition.

I’m a marketing consultant. My computers contain business information and the accumulated work of many years.

Was I really going to install some new AI agent, give it access to my computer, and hope for the best?

Absolutely not.

And this is coming from someone who has spent his entire adult life embracing new technology.

I was excited, but I wasn’t reckless.

My first steps into the water

My early introduction to this new world of agentic AI came through Clawdbot, which eventually became known as OpenClaw.

It was fascinating technology. Open-source software that promised to give artificial intelligence the ability to perform real tasks.

But getting it working was another matter.

I wasn’t about to install it on my primary working computer. So I dug up my old Mac laptop that I could afford to experiment with.

It was short on memory and processing power.

And it struggled.

I spent hours configuring things, troubleshooting problems, taking screenshots, and asking other AI tools how to fix whatever had just gone wrong.

Eventually, I decided I needed more horsepower.

I rented a virtual private server from Hostinger, giving myself more processing power and memory while keeping the experiment isolated from my important systems.

That helped.

But it was still frustrating.

There were configurations to understand, integrations to establish, and APIs to connect. And while OpenClaw itself was open source, the services and AI models needed to make it useful weren’t necessarily free.

I began asking myself whether all this effort was really worth it.

I explored other approaches, including automation platforms like Zapier, Make, and n8n.

I was learning a tremendous amount, but I was also spending an enormous amount of time trying to get technology to do things that I ultimately wanted to accomplish without becoming a programmer.

Then I discovered the Mac mini.

I’d been an Apple user for years, but I’d never owned one of these little machines. I was amazed by its compact size, processing power, and relatively modest price.

I bought one primarily to continue my AI experiments without relying on a rented server.

And suddenly, things started moving.

The performance was remarkable.

But the real turning point was still ahead.

The three words that changed everything

Two AI content creators had caught my attention: Nate B. Jones and Sabrina Ramonov.

Both were discussing developments in artificial intelligence, particularly Claude from Anthropic.

At first, I hadn’t paid much attention to Claude. I associated it with programmers and technically sophisticated users.

Then I heard Sabrina describing a new agentic Claude experience.

Her message, as I remember it, was refreshingly simple.

It just worked.

No endless configuration. No complicated installation process. No weeks of troubleshooting.

It just worked.

After everything I’d been through, those three words were irresistible.

So I tried it.

And you know what?

It worked.

But something else impressed me even more.

Claude kept asking for permission.

Can I do this? Can I access that? Is it okay if I proceed?

At first glance, you might think those interruptions would be annoying.

For me, they were reassuring.

The software wasn’t simply charging ahead, doing whatever it pleased. It was demonstrating boundaries. It was giving me opportunities to review what it wanted to do before allowing it to proceed.

Of course, permission prompts aren’t a guarantee of safety. You still have to understand what you’re approving, and a capable AI system can still make mistakes.

But those safeguards made an enormous psychological difference.

My confidence grew.

I tried something. It worked. I tried something more ambitious. That worked too.

And as the products continued improving, I began trusting them with increasingly meaningful tasks.

Looking back, I realized something important.

What began as a leap of faith had become a progression of earned trust.

The human still has a job

This brings me back to the Tesla.

When I experienced Full Self-Driving (Supervised), I wasn’t surrendering all responsibility to the vehicle.

I was still the driver.

I was watching what the car was doing, assessing the environment, and remaining prepared to take control.

The vehicle handled much of the execution. I remained responsible for supervising it.

Now consider what happens when I use an AI agent on my computer.

I tell it what I want to accomplish. It works through the necessary steps. I observe, approve important actions, and intervene when necessary.

In both cases, the machine is doing work that previously required my direct physical involvement.

And in both cases, my role is changing.

I’m moving from being the person who performs every action to the person who directs, supervises, and makes decisions.

I’ve started calling that role an orchestrator.

It’s a particularly powerful concept for someone running a small business.

I don’t necessarily need to perform every task anymore. I need to understand what needs to be accomplished, choose the appropriate tools, establish boundaries, and evaluate the results.

It’s a form of human-machine collaboration that would have seemed unimaginable just a few years ago.

And I think we’re only beginning to understand its implications.

We’ve been here before. Sort of.

During a conversation about all this with my AI, I found myself thinking about dogs.

Thousands of years ago, humans and wild canids began developing a relationship that eventually led to domestication.

Over many generations, a remarkable partnership emerged.

Humans and dogs possessed different strengths. Dogs could detect things humans couldn’t, alert us to danger, help with hunting, and eventually protect livestock.

We didn’t need dogs to become human.

We learned how to work with another kind of intelligence.

Now, I recognize that artificial intelligence isn’t a dog. It doesn’t possess an animal’s instincts, attachment, or loyalty. And domestication was an evolutionary process spanning generations, not some prehistoric handshake agreement.

But the analogy raises an interesting question.

What happens when humans begin depending on an intelligence that operates differently from our own?

The potential benefits are enormous.

So are some of the fears.

Imagine an early farmer who has come to depend on dogs to protect his flock.

What’s his greatest fear?

Perhaps it isn’t that the dog will fail to protect the sheep.

Perhaps it’s that the dog will turn against them.

And isn’t that remarkably similar to one of our deepest anxieties about artificial intelligence?

We’re not simply afraid that AI might make mistakes.

We’re afraid that something we’ve invited into our lives, something increasingly capable and powerful, might someday act against our interests.

That doesn’t require a machine to develop malicious intentions. Poorly defined objectives, security vulnerabilities, inappropriate permissions, and unpredictable behavior can all produce harmful outcomes.

The lesson isn’t that we should trust AI as though it were a loyal pet.

It’s that powerful partnerships require carefully established boundaries.

Trust doesn’t eliminate the need for supervision. It makes responsible delegation possible.

There’s another leap coming

Today we’re talking about artificial intelligence operating our computers and driving our cars.

But consider what happens next.

Our homes.

Imagine an AI system coordinating your household. Managing energy consumption. Monitoring equipment. Scheduling maintenance. Ordering supplies. Controlling appliances. Working with home security systems. Doing your laundry. Taking the garbage out.

Suddenly, artificial intelligence isn’t merely operating inside a computer.

It’s interacting with the physical world around us.

And once again, we’ll face the same question.

How much control are we prepared to delegate?

I suspect the answer will be the same as it has been throughout my own journey.

Not all at once.

We’ll experiment. We’ll establish boundaries. We’ll observe the results. We’ll become more comfortable as the technology demonstrates its reliability.

And we’ll continue to exercise judgment.

At least, that’s how I believe we should approach it.

You get what you pay for. Sometimes a lot more.

There’s another kind of leap of faith involved in adopting AI that doesn’t get discussed enough.

Opening your wallet.

Every day, I see breathless announcements about some new AI product that’s supposedly going to destroy every established competitor.

It’s free! It’s revolutionary! Everything else is obsolete!

I’ve learned to be skeptical of that kind of hype.

Free and inexpensive AI products can be genuinely useful. Open-source tools can be extraordinarily capable. And paying more certainly doesn’t guarantee reliability or security.

But in my own experience, there’s often a substantial difference between casually experimenting with a free AI service and paying for a capable product that can support serious work.

Today, I’m willing to spend meaningful money on the AI tools I depend on.

Why?

Because I measure their value against what they accomplish, not simply what they cost.

If a subscription costing a few hundred dollars a month can help me perform work that would otherwise consume many hours or require significant outside assistance, the economics can be compelling.

That’s not an argument for buying every shiny new AI product.

It’s an argument for understanding the return on investment.

Sometimes the leap of faith isn’t trusting the technology.

It’s trusting yourself enough to invest in learning what it can do.

The dog at the swimming pool

I want to leave you with another dog analogy.

Picture a young dog standing at the edge of a backyard swimming pool.

Its tail is wagging. It sees its owner enjoying the water. It clearly wants to jump in.

But it’s hesitant.

It’s never done this before.

It doesn’t know whether it’s safe.

Now imagine someone in the pool encouraging it.

Come on! The water’s fine!

That’s how I think about many people who are curious about AI today.

They’re not opposed to it. They’re interested. They can see that something remarkable is happening.

But they’re standing at the edge, wondering whether they should take the plunge.

And here’s where my own experience becomes useful.

Although I’m naturally inclined to embrace new technology, I didn’t jump blindly into the deep end of agentic AI.

I started at the shallow end.

I put a toe in the water.

Then I went down a step.

Then another.

I stood in the shallow end for a while, getting comfortable. Eventually, I ventured into deeper water.

Today, I’m swimming all over the pool.

And I’m having a wonderful time.

Along the way, I’ve introduced AI to business owners and professionals who were curious but uncertain. I’ve watched them discover capabilities they hadn’t imagined were available to them.

The interesting thing is that they don’t necessarily need to repeat all the frustrating experiments I went through.

They can learn from my experience.

They can avoid some of the mistakes.

They can start with tools that already work, establish sensible safeguards, and build confidence much faster.

In a sense, I’ve become the guy in the pool who knows where the steps are.

Maybe even the lifeguard.

Not because I can guarantee that nothing will ever go wrong. Nobody can make that promise.

But because I’ve spent enough time in the water to help someone else approach it more confidently.

And that’s what I think the real leap of faith is about.

It’s not abandoning caution.

It’s not surrendering control.

And it’s certainly not believing every spectacular promise made about artificial intelligence.

It’s having enough curiosity to begin, enough judgment to proceed carefully, and enough confidence to recognize when something genuinely remarkable is happening.

The future isn’t waiting for everyone to feel completely comfortable.

But neither does anyone have to dive into the deep end on their first day.

There are steps.

And for anyone standing at the edge of the pool, wondering whether it’s time to jump in, here’s what I’d say:

Don’t be afraid. Come on in. The water’s fine.