I caught myself doing it again. Three weeks into a new idea and not one line of code written. Just research. Market maps, competitor teardowns, a positioning doc for an audience I had never spoken to. It felt like work. It produced files. It moved nothing.
The research was easy because AI made it easy. I could generate a market map in the time it used to take to open a spreadsheet. So I did, again and again. The cost of the comfortable work dropped to zero, and I optimized straight for it.
the friction gate
There used to be a gate. Deep research was slow, and the slowness asked a question for you: do you actually need this right now? Most of the time the answer was no, and the cost of finding out stopped you. AI removed the gate. The comfortable work is now the frictionless work, and nothing stands between you and an afternoon of it.
That part is new. Procrastination feels like slacking. This feels like progress. That is what makes it worse.
The tell is sequence. Sizing a market before a single line of code. A launch plan for a product that does not run yet. The further the research gets ahead of the build, the more likely it is you are hiding.
fake progress
AI is good at making things that look like work. Slide decks, strategy docs, spreadsheets precise to the decimal. You end the day with artifacts you can point at. None of it touched anything.
That is how the avoidance hides. It wears the costume of productivity. You are not slacking. You are researching. You are planning. You are doing everything except the thing.
the thing
The thing always has the same shape. The cold email you keep redrafting. The call to someone who can say no. A half-built version put in front of a real person. Your name on it, posted, waiting.
What those share is contact. They touch reality, and reality talks back. A model of the market cannot reject you. A user can. A positioning doc will agree with whatever you put in it. The market will not.
AI cannot make the call for you. It cannot build the audience or sit in the discomfort of a rough first version meeting the people it was made for. The work that resists automation is the work that is still yours. It is also the work that can fail, which is the reason you route around it.
settle and build
Research is not the enemy. Sometimes you need more before you can decide. But you are never getting the correct answer, only a good-enough one, and you were always going to settle for that in the end. So settle early and build on it. A week of the product meeting real people teaches you more than a month of analysis ever will.
The analysis loop closes clean and teaches you nothing you did not already suspect. The iteration loop stays open and corrects you the whole way. Choose the one that touches reality.
the tell
When the sensible, well-reasoned, easy work is the work you keep choosing, look at what it is keeping you from. The discomfort is information. It points at the thing that matters, because the thing that matters is the thing that can tell you no.
Follow the discomfort. The spreadsheet was never going to save you.