When you can't start:
how to break an overwhelming task into steps

There is one item on your list. You know what it is, you know why it matters, you even know how to do it. And you have been looking at it for hours.

That is not laziness. How long a task takes and how hard it is to start are two different things, and the second one rarely gets talked about. This page is about the second one: why a big goal locks you up, and how to break it into pieces you can actually begin.

"Finish the thesis" looks like a single line, but it isn't. It hides dozens of decisions: which section, do you need that source first, should you take notes, where is your laptop, should you make coffee.

All of those queue up before you write a word. Your job stops being starting and becomes deciding what to start — and deciding is tiring. The tiredness looks like procrastination from the outside.

This is why "just start" is useless advice. If there is no concrete thing to start, there is nothing to start.

You can do this on paper. The method is:

1. Write the goal, then find the first physical move. Ask yourself: "from exactly where I am sitting, what is the first thing I could do without thinking?" The answer should be as small as "open the laptop".

2. Keep every step to a few minutes. If a step runs past ten minutes it probably has a hidden decision inside it. Split it.

3. Write steps you can observe. A good step is one where you can tell whether it is done. "Understand the topic" is bad; "read three paragraphs" is good.

4. Do not write more than four. A long list becomes a task of its own.

This method has no brand and nobody owns it — it is the shared intuition behind task analysis, chunking and behavioural activation. FocusFlow does it for you, but the method holds with or without the app.

You type the goal exactly as you think of it — "finish the thesis", "start going to the gym", "sort out the tax paperwork". The app splits it into four to six steps, each two to seven minutes long.

The splitting rules are written into the product: the first step has to be a concrete, observable move you can make from where you are, and the steps have to be in the same currency as the goal. If the goal is physical, the steps are physical — it will not hand you desk advice for a gym goal.

Then you can start that first step as a two-minute warm-up. Two minutes is a length that is hard to say no to — and the point was never the two minutes, it is that continuing gets easier once they are behind you.

The AI task splitter is on the free plan. Beyond a daily fair-use limit there is no restriction, and you do not need an account.

Here is the method applied to a real goal. None of the steps say "research" or "plan"; every one of them is a move you can tell you have finished:

By the fourth step you are working on your thesis. At no point did you have to decide to "finish the thesis".

When you use the AI task splitter, the goal title you typed is sent to NVIDIA's AI gateway to generate the micro-steps. Only that title is sent — no account identifier, no email, no other task. We do not otherwise store the prompt or the response.

If the AI is unavailable the app says so and splits the goal using a built-in step list on the device. In that case nothing goes anywhere.

This paragraph says exactly what our privacy policy says. A marketing page that sounds more reassuring than the policy behind it is the worst thing to trust.