TL;DR - My core session is the protected first block of the morning, the most productive hours of my entire day. I write down the task before I start, do the hardest thing first, and stay on one thing until it hits its milestone. When energy dips later, I switch to Pomodoro for a lighter second session instead of pretending I'm still at peak.
Prepare exactly what needs to be done.
The peak. Hardest thing first, one target, agents running in the background.
Don't fight it. Switch gears.
Pomodoro 25:10. Enjoy it, keep moving.
The setup
The first session runs early, in a clean environment with focus audio on and distractions blocked.
- I open my planner.
- I pick the one-to-three things that make the day a win.
- Critically, I write down exactly what this session is for before I touch the keyboard.
Spec first - think before I code
I don't open a file and start typing. I iterate on the spec first and map the edge cases before implementation, because once I start coding my brain flips from thinking into executing and stops finding the edge cases. All the hard thinking goes in while the code is still cheap to change.
One thing at a time
I don't swap tasks until the current one reaches the milestone I set. Hardest thing first, single target, no drifting.
Run agents in the background
- The day before, I prepare what needs doing.
- In the morning, I kick off well-scoped agent work and let it run - autonomously, or while I do the deep thinking on something else.
Background agents on prepared tasks multiply what one block produces.
Two sessions, honestly rated
- The first session is the most productive of my whole day. That's where the hardest work goes.
- When I start to feel it fall off, I don't fight it. I switch to a Pomodoro rhythm (25:10, or 25:5 when I'm less tired) for a second session and let myself enjoy it rather than force full lock-in.
Matching the work to the energy beats pretending the energy is infinite.
Output is the measure
I work 7 days a week. And I try not to burn out - sometimes I deload. 4 hours a day, only the highest-leverage things, then I rest. I'm not the type that grinds long, ineffective hours.
The measure of work is the output you produce. Not the hours. Doesn't matter if it takes a 40-hour week or a 100-hour week - the output is what counts. That's my opinion.
Why it works
- The session pairs the sharpest hours with the highest-leverage work.
- It offloads the mechanical parts to agents.
- It stops when honesty says the peak is over.
It's not about more hours. It's about what happens in the few that matter.