Access isn't capability. Time inside a tool is.

Not long ago, I could name almost every AI tool that hit the market. Not because I researched them professionally, but because I subscribed to them. Writing assistants, image generators, automation platforms, specialized tools for tasks I didn’t even do that often. If someone in a forum said it was worth trying, I was usually trying it within the week.

I told myself I was staying current. Fall behind on the tools, fall behind on everything. That’s what I kept telling myself, anyway.

What I was actually doing was spending over $200 a month on subscriptions I didn’t have time to use, building a stack I couldn’t manage, and calling it productivity.

It took me longer than it should have to see that. And when I finally did, the fix wasn’t finding better tools. It was getting honest about the ones I already had.

The one that still makes me shake my head a little is Zimmwriter.

Zimmwriter is a writing tool with a solid reputation in certain corners of the online marketing world. People were using it to produce content at scale, and the lifetime deal made the math look easy. Pay once, use it forever. No monthly subscription eating into your overhead. I bought it without hesitation.

I used it for a hot minute and moved on, telling myself I’d come back.

Not because it was a bad tool. From what I could tell in that brief window, it did what it promised. But learning it properly meant carving out real time. Building prompts that fit my workflow. Running it through enough repetitions to know whether it was actually saving me anything.

I never did any of that. Something more urgent always came up, or something newer caught my attention, and Zimmwriter sat quietly in my browser bookmarks while I moved on to the next thing.

That’s the part nobody talks about when they share their AI stack. The tools that are technically still active. The trials that converted to paid because canceling felt like admitting defeat. The lifetime deal that was only a deal if you actually used it.

I wasn’t the only one doing this. I just finally stopped pretending I wasn’t.

At some point I actually sat down and added it up.

ChatGPT, Claude, Gemini, TypingMind, image generation tools, a couple of automation platforms, and a handful of others I’d signed up for and honestly half-forgotten about. Over $200 a month.

When I saw that number, my first reaction wasn’t outrage. It was embarrassment. Because I knew, if I was being straight with myself, that I was regularly using maybe two of them.

The rest were just sitting there. Billing me.

I’d told myself the same thing about each one: I’ll get back to it when things slow down. Things never slowed down. And I never got back to them. I just kept adding new ones and pushing the half-learned ones further down the list.

What I had wasn’t a stack. It was a graveyard with a monthly subscription attached.

So I started canceling. Not all at once, not with some big strategy behind it. I just stopped renewing things I couldn’t justify, one at a time, until I got down to the one tool I actually opened every single day.

That was ChatGPT. Everything else went.

When I canceled everything except ChatGPT, I wasn’t making a grand declaration. I was just tired of paying for things I wasn’t using.

But something unexpected happened when everything else was gone. I actually started using the one tool I had left. Not dipping in and out of it between sessions with four other platforms. Not comparing it to something I’d read about that week. Just opening it, working in it, and figuring out what it could actually do.

That sounds obvious. It wasn’t obvious to me at the time.

What I found was that a tool you use every day for a year becomes something different than a tool you use occasionally. The first few months still felt like work: figuring out how to prompt it well, learning where it was strong, and where it needed more direction from me.

There were plenty of sessions where I walked away thinking I could have written that faster myself.

But somewhere around month three or four, something shifted. I stopped thinking about how to use it and started just using it. The workflow stopped feeling like a process I was following and started feeling like a rhythm. That’s a different thing entirely, and it doesn’t happen in a free trial.

I’m not saying a year is the magic number for everyone. But I am saying that most people quit long before they find out what their tool can actually do. They hit the learning curve, decide the results aren’t good enough, and go looking for something better. The something better has the same learning curve. And the cycle starts over.

I know because I ran that cycle for a long time before I got off it.

People ask me sometimes how you know when you’ve actually gotten good at an AI tool. It’s a fair question, and I didn’t have a clean answer for a while.

Here’s what I’ve landed on.

You know you’ve gotten there when you stop thinking about the tool and start thinking about the work. When you open it the same way you’d open a notebook, not because you’re excited about the notebook, but because that’s where the thinking happens. The tool becomes infrastructure.
It stops being the interesting part.

There’s also a more practical signal. You start hitting the tool’s actual limits instead of your own. Early on, when results disappoint you, it’s usually because you don’t know how to ask well. You’re working around your own gaps, not the tool’s. When you’ve genuinely gone deep, the frustrations change character.

You know exactly what you’re asking for, you’re asking for it clearly, and the tool still can’t quite get there. That’s a real ceiling. That’s when it might be worth looking at something else, not before.
Most people never reach that point. They mistake their own learning curve for the tool’s limitation and move on. I did this more times than I’d like to admit before I finally recognized the pattern.

The other thing that happens when you’ve gone deep enough is that you start seeing possibilities the demos never showed you. The use cases you find through daily use are almost always more specific and more useful than anything in a tutorial.

They come from your actual workflow, your actual problems, your actual voice. Nobody else’s demo will surface those for you. Only time inside the tool will.

For a good stretch, ChatGPT was it. One tool, one workflow, nothing else.

Then a couple of months ago I added Claude back in.

Not because I was bored with what I had, or because someone told me Claude was better, or because I saw a demo that made me feel like I was missing out. I added it back because I had a specific need that ChatGPT wasn’t covering as well as I wanted, and I knew Claude well enough from earlier use to believe it would cover it better.

That’s a different reason than the ones I used to subscribe for things.

Before, I added tools because they looked promising. Because the price was right. Because someone in a forum was enthusiastic about it, or I’d just finished a training that made it sound essential. Those are feelings masquerading as reasons.

This time, I had a real gap in a real workflow, and a tool I already understood well enough to know it could fill it.

That’s the bar I use now. Not “does this look useful.” Not “could I see myself using this someday.” The question is whether I have a real, specific need that my current tools genuinely can’t meet, and whether the new tool has earned enough of my trust to get a real look.

Most of the time the answer is no. And that’s fine.

Right now I have two tools. I open both of them every day. I know what each one does better than the other, and I use them accordingly. That’s the whole stack. It costs less than what I was spending on tools I’d forgotten I had, and it does more.

I’m not going to tell you which tools to use. That’s not really the point of this.

What I will suggest is one exercise that changed how I thought about all of it. Go look at every AI subscription you’re currently paying for. Not the ones you’re planning to use. The ones you actually open regularly. Be honest with yourself about the difference.

For most people, that list is shorter than they think.

Then ask one question about each tool that didn’t make the cut: have I given this a real run, or have I just given it a start? There’s a big difference between a tool that genuinely didn’t work for you after sustained use and one you opened twice and set aside. One of those is a decision.

The other is just an unfinished experiment with a monthly fee attached.

If it’s unfinished, you’ve got two choices. Give it a real thirty days and find out what it can actually do. Or cancel it and stop pretending you’ll get back to it.

Neither of those is the wrong answer. What’s costly is the middle ground, keeping something around because canceling feels like giving up, while never using it enough to know whether it was worth keeping.

That middle ground is where most of my $200 a month was going.

It doesn’t have to be dramatic. Take one honest look at what you’re actually using, decide what stays and what goes, and you’ll have a cleaner stack and a clearer head before the week is out.

I started this piece by telling you I was spending over $200 a month on AI tools I didn’t have time to use. That wasn’t an AI problem. It wasn’t even really a tools problem.

It was a focus problem dressed up as a strategy.

The thing I keep coming back to is that none of this is new. Long before AI was part of the picture, I watched the same pattern play out with marketing platforms, email tools, and every other category of software that promised to make the business easier. The tool changes. The habit of collecting without learning doesn’t change on its own. You have to change it deliberately.

What changed for me wasn’t finding better tools. It was deciding that understanding one thing deeply was worth more than having access to everything.

That’s still true. It was true before AI made the options infinite, and it’s true now that they are. The founders getting real results from these tools aren’t the ones with the longest subscription list. They’re the ones who stayed with something long enough to find out what it could actually do.

Two tools. Used every day. That’s the whole stack.

It’s enough.

If my story sounds familiar, the newsletter is where I write about what I've learned from it. Above the Noise goes out every week, honest thinking about trust, attention, and building something real online. No hype, no shortcuts.

Subscribe for free on Substack

David Wakeman
Operate above the noise

© Copyright David Wakeman.  All Rights Reserved