I caught myself the other day.
Not for the first time. But clearly enough that I couldn't ignore it.
I was sitting there, deep in another reset. New plan. New structure. New clarity about how everything was going to work.
And it felt great. Clean. Focused. Almost like I'd stepped into a better version of my life.
Then it hit me. I've done this before.
Not once. Not occasionally.
Constantly.
The Comfort of Starting Again
There's a particular feeling that comes with a new beginning. Something lifts. The old version of the project, with all its wrong turns, accumulated compromises, structural problems that revealed themselves too late, gets set aside. A new version opens. It knows what the old one didn't.
The seductiveness of this isn't irrational. Starts genuinely are different from continuations. The energy is different. The clarity is real. You can see the whole shape of a thing before the complexity has had time to settle in.
The problem isn't the instinct. The instinct is often right. The problem is the pattern: when starting again becomes the thing you do instead of the other thing. The harder thing. The staying.
AI Changes the Game
Before, ideas had weight.
To act on them meant time, people, energy. You had to push them into the world, and that friction filtered things out. Most ideas stayed as thoughts.
Now, there's no gap. You can build a system in minutes. Prototype something straight away. Turn a vague instinct into something that looks finished.
It's incredible. Honestly, I'm doing things now I couldn't have imagined a few years ago. There are moments where it genuinely feels like stepping into a different universe.
And I love it.
The Speed of Change
The pace of the technology itself is dizzying. You start to learn a tool, build a workflow, get some traction... and then something new appears that's clearly better.
So you switch. You unpick what you've just built. You optimise again. Which feeds the same instinct:
This isn't quite right yet. There's a better version. Just reset.
The New Narrative
I came across a video recently that framed this in a very different way. The argument was simple:
What we've been calling ADHD might actually be an advantage in the age of AI.
The reasoning makes sense.
- People who think across domains
- People who connect patterns quickly
- People who don't naturally stay in one lane
These are the people who tend to generate new ideas.
I found that genuinely compelling. Not as an excuse, but as a reframe. The multi-lane brain isn't broken. It just needs the right conditions. And AI, finally, might be those conditions.
Where This Actually Works
I can feel it working. Genuinely.
Fifteen projects running in parallel: a film score, a solo album, an immersive installation concept, two writing projects, a software prototype. And I can move between them in a way I never could before. Build context fast. Pick up a thread I set down three weeks ago. Carry multiple worlds simultaneously.
The ideas that previously stayed in notebooks because the cost of starting them was too high are now getting made. That's real.
And in the spaces between projects, I'm connecting things I wouldn't have connected otherwise. Threads that belong together but lived in separate categories. The associative thinking that used to scatter me is suddenly a feature.
But It Doesn't Solve Everything
Here's where I need to be honest.
Moving fast between things isn't the same as finishing things. Starting again feels productive. Often it genuinely is. But there's a specific point, and I've started to recognise it, where the reset becomes something else.
The new project is always better than the old one, in a particular way: the new project hasn't been tested yet. It hasn't revealed its problems. It still has all its original energy and none of its consequences.
That's not a feature. That's the trap.
The Loop
This is the thing I've been sitting with.
The instinct to start again is real and often useful. The acceleration from AI is real and often extraordinary. But together they create a specific kind of loop: one that feels like progress because it always begins with clarity and energy.
You start. You build. You get traction. Then the novelty fades, the difficulty appears, or a new possibility emerges. And the instinct fires: reset.
Because resetting is now so fast, so frictionless, there's almost nothing to stop you.
And so you do it again.
The Addiction
I think this is something closer to an addiction than a productivity problem.
Not to the work itself. To the beginning of the work.
That initial clarity. The sense of having finally understood something. The version of yourself who knows exactly where this is going and why it will work this time.
The beginning always feels better than the middle. In the age of AI, you can always be at the beginning.
What Actually Helps
I've tried the systems. The frameworks. The productivity architectures that promise to capture the loop and convert it into forward motion. They don't work. Or they do until I reset those too.
What actually interrupts the pattern is simpler and harder: constraint.
Specifically, finishing something that doesn't feel right. Shipping the version with flaws. Publishing before I've resolved the part that still bothers me.
The constraint forces the middle to become real. And real things can be finished. New plans, by definition, cannot.
Using AI Differently
What I'm learning, slowly, is that AI is most useful precisely where the loop is strongest.
Not at the beginning. I don't need help beginning. The beginning is the part that's easiest for me.
Where I need help is in the middle. The sustained, grinding, unsexy part of making something exist. The part where the novelty has gone and what remains is just the work.
That's where the tool actually earns its place. Not as a better beginning, but as a way through.
Thinking More, or Thinking Less?
There's a lot of anxiety around AI making us mentally lazy.
And I understand that.
If you use it to avoid thinking, it probably will. If you ask it to write everything, summarise everything, decide everything, then yes, something is being offloaded.
Maybe too much.
But that hasn't been my experience.
If anything, I feel more mentally active.
Not because AI gives me answers.
Because it gives me a surface to think against.
I can follow a question instantly. Test an idea. Compare two possibilities. Find the edge of a thought and push it a bit further.
It's not replacing the work of thinking.
It's accelerating the movement between thoughts.
And maybe that's where the real difference lies.
AI can be a shortcut around thought.
Or it can be a catalyst for it.
The danger is mistaking one for the other.
The Ulysses Contract
I like the idea of a Ulysses contract.
A decision you make while clear-headed, to protect yourself from the version of you that will get carried away later.
That's what my North Star document is becoming.
Not a productivity system. Not another framework.
A rope around the mast.
Because the siren song now isn't just distraction.
It's reinvention.
The promise that if I start again, everything will finally align.
So I need something fixed.
Something written when I'm calm enough to know what I actually want.
Something that can say:
Not today.
Keep going.
Where I've Landed
I haven't solved this. I don't think it's solvable. It's a pattern that gets managed, sometimes well, sometimes not.
What's changed is the awareness. I can see the loop now. Sometimes I catch it early. Sometimes I'm three weeks into a new plan before I recognise it.
But I've started to notice something. The thing I'm resetting from, the difficult, stuck, not-quite-working version of the project, is usually closer to finished than the thing I'm resetting to.
The loop is a map. And the exit is usually behind you.
You don't just move faster.
You spin.