Brain Frizz, Part 1: Blocking My Day
I review a lot of pull requests (PRs) now. Some of that is the job. As a staff engineer, part of my work is helping other people get unstuck, and on a team where AI agents write a lot of the code, what's usually stuck is a PR waiting on a review.
The rest of my day goes to my own tickets, which usually means a few agent sessions running at once. Between the two, I was switching all day: start a ticket, review a PR, check on an agent, answer a ping, then go back to the ticket and try to remember where I was.
I call what that does to me brain frizz. It's a foggy, headachy feeling where I can't concentrate on one more thing. The tension in my shoulders creeps up, and I'm more irritated at everything. It was at its worst the afternoon of November 6, which is when I sat down and planned this experiment.
So for two weeks, November 11 to 24, I'm splitting my day into fixed blocks, some for ticket work and some for reviews. This post is the plan. I'm publishing it at the halfway point, and I'll write up the results after the 24th.
Why Blocks
I've been digging into our team's PR history for a post on WIP limits (WIP means work in progress). The short version: we open a lot more PRs than we did two years ago, they're much bigger, and the wait for a first review has grown several times over. Review is where our work piles up now.
A WIP limit caps how many things are in progress at once. Blocks are my personal version: each part of the day holds one kind of work, so reviews and my own tickets are never in progress at the same time.
What My Calendar Showed
Before picking times, I went back through two months of my calendar. Meetings almost always landed between 10:30 and 12:30. Afternoons were mostly open.
So the blocks go around that cluster instead of competing with it.
The Schedule
| Time | Block |
|---|---|
| 9:00-10:30 | Ticket work |
| 10:30-12:30 | Meetings, when there are any |
| 1:00-2:00 | Reviews |
| 2:00-4:00 | Ticket work |
| 4:00-4:45 | Reviews |
They're recurring events on my calendar every weekday (Monday's morning block starts at 9:15), and I told the team about them. That helps less than it sounds. Nobody checks a calendar before sending a Slack message, and review requests go to a shared channel where everyone posts their PRs. So holding the blocks is up to me. A review request that comes in during a ticket block waits for the next review block, a few hours at most during the day.
The exception is an urgent review: something blocking a deploy, or a teammate who can't move until I look. Unblocking people is part of my job, so those can break a ticket block. I still log them as broken blocks, so I can see how often it happens.
What I'm Tracking
I check in twice. At 8:50, before the first ticket block, I score my brain frizz as fog from 1 (clear) to 5 (fried) and do a body check: headache, tight shoulders, irritability, and how much work followed me into the evening. At the end of the day, a two-minute log covers the rest:
- Whether a ticket block broke, and what broke it: a meeting, someone else (a ping or an urgent review), or me deciding to switch
- How many PRs I reviewed, and how many review requests didn't fit in a review block
- The most agent sessions I had running at once during a ticket block
- How ticket work felt: real progress, mixed, or pushed around
The broken-block causes are the part I care about most. If most of the breaks are me choosing to switch, a new calendar won't fix much.
My baseline is November 10, the day before the experiment started. By then the worst had eased to some mild fog, but I still had tight shoulders and reviews sneaking into my ticket work.
What Counts as Success
I'm writing this down now so I can't move the goalposts in the results post.
It worked if my average fog score for the second week is at least half a point lower than my November 10 score, with no more than two days at 4 or 5.
It failed if fog doesn't improve, or if teammates start waiting more than a day for my reviews because of the blocks. Feeling better doesn't count if the cost lands on everyone else.
Halfway Through
The first week, I didn't fill in the log. I'd built it before the experiment started, so it was there. But I put it somewhere that never came up during my day, and filling in a log at 4:45 isn't a habit for me yet. So I didn't do it.
I've moved it next to the daily journal I already open every day, and I'm logging from today on. The results will lean on the second week.
Even without the log, I'm feeling better. The part I didn't expect is the conversations. Once I started talking about the blocks, a lot of my coworkers said they're feeling the same thing, and each of them has a different fix. Some are leaning into AI as the fix for AI, building dashboards that show where their attention should go next. Others are blocking time too, or at least starting the day with reviews. That's good. What works depends on the person and their workflow, so try what fits yours instead of copying my schedule.
I'll post the results after November 24.


