Claude Code Is My Operating System. Here’s What That Actually Means.

The Sunday Night Honesty Check

It’s 10:47 PM on Sunday, and I’m sitting with a 5.4/10 day rating that feels exactly right. Not bad. Not good. Real.

I spent the afternoon in “Prep 4 Tmrw” mode—which I actually did well. Loaded my custom GPT stack: Baby Love for parenting angles, Kyle for career moves, Dave R for financial architecture, Ben for news context. Built Granola folders with Dana, Apollo, ATC, and work intel. That’s the *right* kind of Sunday prep. I wasn’t spinning; I was actually thinking about Monday’s shape.

Portfolio came through clean. VST took a 6.7% hit, GEV down 7.0%—stings. But ASML +3.1%, NVDA +3.6%, AMD +8.0%—enough green to keep the Leopold Aschenbrenner AI recursion thesis alive. I’m tracking without panicking. That’s discipline.

And then Daphna messaged about Lior’s sleep—front-door logistics, the usual Sunday evening real-life bleed. But we’re *naming* it in couples coaching, working through it instead of ignoring it. That matters more than pretending it’s smooth.

Where I Failed—And Why It Matters

I didn’t work out. Ten-day streak, broken. Strava errored (HTTP 400), but let’s be real: I didn’t *make time* for it on a Sunday when I theoretically had the most control. That’s a miss I’m owning.

Worse: my automation stack is fractured.

Tomer Morning errored out. So did Morning Email, ScreenPipe Brief, Blog Generator, Budget Updater. I’m setting up infrastructure (ChatGPT desktop installer visible in Finder, three versions of it sitting there), but I’m not *maintaining* systems that already exist. Classic pattern: build → abandon → rebuild.

I have 11 incomplete tasks in Skilled Hunters, 213 in Marketing, 100 in Weekly Priorities. Not alarming numbers. But the pattern is clear: I’m adding complexity without closure.

The real problem? I’m prepping *around* broken systems instead of *fixing* them. I curated news like I’m building intellectual capital, synced with Daphna on parenting—all good. But if Tomer Morning is erroring and my automations aren’t running, I’m losing signal. I’m optimizing the visible layers while the foundation cracks.

That’s not startup energy. That’s unmanaged technical debt.

Three Things I’m Fixing Monday Morning

Debug Tomer automation stack: The morning brief error, then re-enable the five stalled jobs. Verify Cron health. This runs while I sleep or sells—it has to work.
Fix the streak: Workout Monday morning. No excuses.
Clear Alfred bleeding edge: Lior admin task, automation enablement, duplicate “Add Chrd” tasks. Get these off my mental RAM.

The Question I’m Actually Sitting With

Here’s what Sunday night taught me: I can *prepare* at a high level. I can think about career moves, portfolio positions, parenting dynamics, automation architecture all at the same time. That’s real.

But I’m also someone who breaks fitness streaks on Sundays and leaves automation systems in error state without fixing them immediately. I’m a sales leader who builds agents on the side, and that tension means something has to give—and right now, it’s usually the infrastructure.

So the real question: Is broken automation infrastructure a feature of how I operate (fast-moving, tolerant of debt), or am I actually just procrastinating on maintenance because it’s not as visible as prepping for tomorrow?

Yalla. Monday will tell me.