I've always hated the idea of short work sessions. trydeepwork was built around focus, and to me, focus meant sitting down for at least half an hour. Can you really call it focus if it only lasts five minutes? Fifteen? I never thought so (and still don't!)

I built the whole levels system around this belief - anything less than thirty minutes always felt like cheating, so I avoided building it into trydeepwork. 30 mins to an hour was just getting started (L0). Real focus sessions started at an hour (L1), got serious at two hours (L2), and anything past four hours earned the top tier (L3).

But as I started building the tool, I began noticing my own habits more closely. And the biggest challenge wasn't completing a two-hour coding session, it was sitting down to get started in the first place.

Starting a task feels like trying to push a dead car. Once it's moving, you can keep it rolling, but that first shove is just brutal. Even cars need a starter motor just to get going. That first shove - what researchers call the "task initiation problem" - turned out to be my real enemy, not the length of focus sessions.

Enter microWork (μW)

So after months of resisting, I eventually caved and built microWork (μW): sessions of 25 minutes or less, starting from just 5 minutes.

These aren't tracked like regular sessions. They don't count toward your deep work hours, don't use up your free session limits, and don't get special analytics treatment. They're utility sessions - designed purely to help you get started or knock out small tasks, not to be celebrated or optimized.

Think of them as a starter motor: sometimes you only need the smallest excuse to begin, and you can always extend them into actual focus time (30+ minutes) when momentum kicks in.

Most of my "10-minute" sessions end up stretching well beyond 10 minutes, but sometimes they don't - and that's fine too. At least I nudged the car an inch. Turns out the hardest part wasn't building better focus - it was just getting the engine to turn over in the first place.