There’s a version of this post that starts with a grand claim about attention being the new oil or something equally exhausting. I’m going to skip that. What I actually want to talk about is something quieter and harder to pin down: the way the tools you use every day gradually reshape what you notice, what you remember, and what kinds of thoughts you end up having at all.
What makes it hard to notice is that the change usually happens quietly, through repetition, until a tool starts to feel less like something you use and more like the room your thinking lives inside.
The environment shapes the thinking
When I started getting serious about building systems for my work, the initial draw was simple: I wanted my goals and tasks somewhere concrete instead of constantly juggling them in my head. That still makes sense. Externalizing mental load is a real thing with real benefits. Write it down, free up the RAM, move on.
But there’s a second-order effect that took longer to notice. When you consistently offload the same kinds of thinking to the same kinds of tools, you start doing less of that thinking on your own. The tool doesn’t just store the thought. It starts replacing the act of forming it.
This is not a catastrophe, but it is worth being honest about: the environment shapes cognition, and the question is whether you’ve chosen that environment with any intention.
What note-taking apps actually do to memory
Most note-taking apps are sold on capture. Get it out of your head, into the system, searchable forever. And that’s useful. But there’s a cost buried in the design.
Memory isn’t just storage. It’s a process. When you sit with something long enough to remember it, you’re doing work that changes how you understand it. You’re connecting it to other things you know. You’re noticing what still feels unresolved. You’re, in a small way, thinking.
When capture is frictionless enough, you skip most of that. The note exists. You feel like you’ve handled it. But a week later, when you could use that idea, it’s sitting in a folder you’ve already forgotten about, searchable in theory, inaccessible in practice.
I’ve watched this happen with my own notes. The density of what I’ve captured and the density of what I’ve actually integrated into how I think are not the same thing, and the system can become a place where ideas go to be filed away rather than a place where they get worked on.
Feeds and the attention they train
Feed-based platforms are designed around a specific model of attention: short, reactive, continuous. You scroll, something catches you, you react or move on, you scroll again. Repeat until you have to do something else.
The issue isn’t that this is always bad. The issue is that the more time you spend in that mode, the more you train your attention toward short arcs, fast resolution, and low tolerance for ambiguity or slow build.
And then you sit down to think through something that actually requires sitting with it for a while, and it’s harder than it should be. Not because you’ve become less intelligent, but because you’ve been practicing a different skill. Attention is trainable in both directions.
I find this genuinely concerning, not in a moralizing way, just in a practical one. Some of the work I care about most requires a kind of sustained, wandering, slightly uncomfortable attention that feeds are not built to support. They’re built to interrupt it.
Automation and the distance problem
There was a period where I tried to automate a lot more than I do now. Workflows firing in multiple directions, AI handling first drafts of responses, systems practically running themselves. I was proud of the architecture. It was clever and it saved time.
But something started to feel off. The distance between me and the actual work had grown. I wasn’t making decisions so much as reviewing outputs. And some of what I’d automated were things that, it turned out, I needed to do myself in order to understand what was actually going on.
There’s real value in automation, and I’m not reversing that position. But the specific value is freeing up attention for things that require it, not outsourcing judgment in places where exercising judgment is the whole point. Those two things can sound similar, but they lead to very different systems.
When I automate something, I now ask: is this reducing noise so I can think better, or is this creating distance from something I actually need to stay close to? The answer changes what I decide to hand off.
AI tools and the question of whose thinking this is
AI writing and thinking tools make this sharper. They’re genuinely capable and I use them, but the cognitive dynamic is worth looking at clearly.
If you prompt an AI to draft your thinking for you before you’ve done the thinking, what you get back is a fluent version of a thought you didn’t quite have yet. It may look like your position and read well on the page, but you didn’t arrive at it so much as receive it. That’s a meaningful difference, especially if the work is supposed to express something you actually understand or believe.
This isn’t an argument against using AI tools. It’s an argument for sequencing. Work out what you actually think first, even roughly, even messily. Then use the tool to extend or sharpen it. That order matters. Reverse it often enough and you start to lose track of where your thinking ends and the tool’s begins.
I care about this partly for practical reasons (the work is better when it’s actually grounded in something I’ve thought through) and partly for something harder to name. There’s a kind of self-respect in doing your own thinking. In keeping your voice intact. In not outsourcing the part of the work that is actually you.
Choosing your cognitive environment
None of this is an argument for using fewer tools or simpler ones. It’s an argument for using them with some awareness of what they’re doing to how you think, not just what they’re doing to your workflow.
The tools you spend the most time in are teaching you something. The feed is teaching your attention something. The note-taking app is teaching your memory something. The automation stack is teaching your judgment something. The AI assistant is teaching your relationship to your own ideas something.
That is not inherently bad, but it is happening whether you’ve thought about it or not.
What I’ve come to believe, through more trial and error than I’d like to admit, is that the goal isn’t a frictionless system. It’s a system that keeps you in contact with your own thinking. One that handles the noise without handling the substance. One that creates more room for the kind of attention you actually want to develop, rather than quietly training it out of you.
It is a harder bar to design for, but it feels like the right one.
Digital Zen 