I was three paragraphs into a piece I was writing when I opened a new tab to look something up. By the time I came back, I had started outlining a different post, sketched the rough idea for a third, and was halfway through asking an AI to summarize a book that I had never read. None of this took more than a few minutes. All of it felt like work.
It wasn't until I closed everything and looked at what I had actually finished that I noticed: nothing.
I had been busy. I had been generating. I had been starting things. But the original post was still paragraphs long, and I now had four half-formed ideas sitting in front of me instead of one finished one.
This has been happening a lot lately. And I'm starting to realize that it may not be a discipline problem. This might be something that the technology itself is doing.
The barrier to starting something has vanished. Whether that's a good thing isn't obvious.
Before the current wave of AI, accomplishing something took real work. Time felt more limited. If I wanted to learn a new skill, I had to spend hours just figuring out where to start. I had to search for the right book, the right tutorial, the right teacher. And this is assuming I managed to stay focused and refrained from forming new ideas to expand on. The information existed, but it was scattered, and pulling it together was its own kind of labor. Skills were bottlenecks, and they were not always easy to clear.
Now, that barrier has almost entirely disappeared. If I want to write, build, or learn something, I can begin almost instantly. I don't need to find the right resources. I can just ask. The friction that used to slow me down and force me to choose carefully is no longer there.
At first, this may sound like a clear improvement, but I'm not so sure anymore.
Constraints used to do quiet work in the background. They didn't only limit what I could do; they helped decide what I actually did. When not everything was possible, I was forced to prioritize without thinking of it as prioritizing. The world made the cut for me.
Now, nearly everything feels possible. But my capacity remains unchanged. I still have the same amount of hours in a day, the same attention, the same finite energy. The difference is that I now have access to far more than I can realistically act on, and the gap between what I could do and what I can do has stretched wider than ever.
This creates a new problem, and a quietly difficult one. I have to decide what not to do.
That sounds simple. It isn't.
Choosing what not to do is a different cognitive skill than choosing what to do, and I don't think it is as widely developed yet. For most of human history, it didn't need to be. The constraints did that work. You wanted to start a business, but you didn't have capital, so you didn't start it. You wanted to learn an instrument, but you couldn't afford lessons, so you didn't learn. The unchosen options didn't feel like decisions. They felt like reality.
When reality stops doing the filtering, we have to do it ourselves. We're not very good at it.
I keep coming back to this distinction because it matters more than it first appears. AI doesn't just make tasks easier. It changes what a decision is. In a world of high friction, deciding to do something was the end of a long process of weighing, planning, and committing. The decision came after the deliberation. In a world of near-zero friction, deciding to do something is itself the beginning. You can start now and figure out later whether it was worth starting.
The strange part is that the drifting feels productive.
I enjoy learning and exploring new ideas. Whether I'm researching or starting something new, the experience produces the same internal signal: that something is happening, that progress is being made. But I've started to suspect that I'm being trained to do what feels good, not to move toward anything in particular. The dopamine of novelty is indistinguishable, internally, from the satisfaction of finishing.
This isn't entirely new. Phones already trained us to crave switching. What AI changes is the category of activity that switching applies to. Scrolling Instagram felt like a distraction even when I was doing it, because I knew the activity wasn't producing anything. Opening a new tab to brainstorm a different project feels like work. The output looks like the output of effort. Words appear. Drafts accumulate. Ideas get generated. The signal of productivity is preserved even when the substance isn't.
Social media made distraction feel like rest. AI makes distraction feel like work.
That distinction is what I keep getting caught on. Distraction that feels like rest is at least honest about what it is. Distraction that feels like work is harder to notice, and much harder to interrupt, because nothing inside the experience tells you to stop. You aren't zoning out. You are generating. You are producing artifacts that look exactly like the artifacts of someone making real progress. The disguise is nearly perfect.
There's an old idea that starting is the hardest part. Once you begin, momentum carries you. That was probably true when starting required effort. The act of beginning was a kind of commitment in itself, because of everything it cost to get there. By the time you'd organized the resources, cleared the time, and overcome the inertia, you had already invested enough that staying with it was the natural choice.
The equation has flipped. Starting now costs essentially nothing. This means starting no longer functions as commitment. It functions as exploration, which is a different thing entirely, and not always a useful one.
I don't believe the answer is to manufacture artificial constraints, though I've seen people try. The "use AI less" approach treats the problem as a discipline issue. I don't think it is. The problem is that abundance has exposed a skill we never had to develop. We need a different relationship with our own attention. One that doesn't depend on the world doing the choosing for us.
What that looks like in practice, I'm still figuring out. I think part of it is being willing to let promising things go unpursued. Not because they aren't worthwhile, but because pursuing them costs something even when AI makes starting free. Attention is still finite. Follow-through is still finite. The fact that I could learn five new things this month doesn't change the fact that I can only really commit to one or two.
The other part is being honest with myself about what I'm actually doing when I start something new. Am I beginning something I intend to continue, or am I enjoying the feeling of beginning? Those are different activities, and I think AI has made them harder to tell apart.
I notice I'm still doing more than ever. I'm just not sure I'm moving toward anything.
I don't have an answer. I'm not sure there is one yet. But it seems like a question worth sitting with, especially because so much of what AI offers makes it easy to skip sitting with anything at all.