The Vacation Velocity Experiment
$87,150 in created value. 92 commits. 19 days.
That was my vacation. I don't know if that's inspiring or depressing... probably both.
I took three weeks off in January. No meetings. No Slack. No "quick calls" that eat 45 minutes and produce nothing. Just me, my codebase, and AI tools that have gotten genuinely scary good. I wanted to answer one question: what happens when you remove the job from the developer?
Not "what happens when a developer works on a side project." What happens when the 6-8 hours of daily cage time just... disappears?
Week 1 was humbling. 10 commits across 7 days. Garbage. I was grinding through what I'm calling the decompression tax -- your brain doesn't just switch from employee mode to operator mode because you closed your laptop. It took me a full week to stop checking email, stop context-switching between three imaginary projects, stop writing code defensively like someone's going to "review" it next sprint. By day 7 that gravity finally broke... and then everything changed.
Week 2 I started shipping real shit.
| Date | Value | What Shipped |
|---|---|---|
| Jan 8 | $3,450 | Identity system upgrade |
| Jan 9 | $3,225 | Payments and checkout UX |
| Jan 11 | $3,300 | OS-style navigation |
| Jan 12 | $5,100 | Public metrics API, homepage |
| Jan 13 | $9,300 | Finance dashboard sprint |
| Jan 14 | $4,650 | Content engine sprint |
$29,025 across 6 active days. $4,800/day average. Identity systems. Payment flows. Navigation. Dashboards. Not prototypes -- production code deployed to Cloudflare Workers, serving real requests. January 13 was the breakout. 12 commits. 7 before noon. $9,300 in a single day... which is more than my biweekly paycheck lol.
But yeah... week 2 was just the warmup.
Week 3 the compounding kicked in and it got kind of absurd.
| Date | Value | What Shipped |
|---|---|---|
| Jan 15 | $9,600 | Documentation, tech dashboard, assessment |
| Jan 17 | $16,050 | Freedom and fitness domains |
| Jan 18 | $15,975 | Finance Nexus production ready |
| Jan 19 | $16,500 | Strategic pivot, Finance Nexus phases D/E/F |
$58,125. Five days. $11,625/day average. I was speedrunning an entire SaaS platform solo and the curve was still going up.
2.4x the output of week 2. Not because I worked harder. Because everything I built in week 2 was now infrastructure. Components composed. Patterns repeated. The codebase had its own momentum. I was basically min/maxing every hour -- stack the right abstractions early, reap the compound interest later. File size is an architecture decision and those early splits paid for themselves in week 3. New game+ energy except the game was real.
I don't know... looking at these numbers even now it sounds insane. $16,500 in a day? From one developer? Even my shit is gold apparently... but the data's right there in the git log. Every commit timestamped. Every feature deployed.
Either way, here's the part that actually matters.
January 20. First day back at work. Two commits. $2,400 in value. A maintenance day. UX polish and dependency updates. The math:
| Metric | Vacation | Regular Work | Drop |
|---|---|---|---|
| Hours per day | 8-10 | 1-2 | 5-8x less |
| Commits per session | 9+ | 2-4 | ~3x less |
| Value per session | $8,700+ | $1,500-3,000 | 3-5x less |
One day back and my output dropped 3-5x. Not because I forgot how to code. Because the job reclaimed 6-8 hours of my day, plus the mental overhead of context-switching between someone else's problems and my own. The cage doesn't just take your time. It takes your momentum.
The most important finding wasn't the peak output though. It was the curve. Week 1 was almost nothing -- if I'd measured after 7 days I would've concluded the experiment failed. The decompression tax is real and unavoidable. Your brain needs time to respawn in a different mode.
Week 2 was $4,800/day. Strong.
Week 3 was $11,625/day. Absurd. And still accelerating.
That's the part that fucks with me. The curve hadn't flattened. I didn't hit a wall. I hit a return-to-work date. I ran out of vacation before I ran out of velocity... and I don't know what week 4 would've looked like but the trendline says something I'm not ready to ignore.
Regardless... the data's all there. $87,150 in value. 92 commits. 10 active coding days. One developer. No team. No sprint planning. No standups. No retros. Then I went back to work and produced $2,400 in a day.
My employer pays me $92,000 a year for my time. In 19 days of self-directed work, I created value worth 94.6% of that annual salary. That's the data behind freedom being a math problem... the numbers don't lie.
The job isn't the runway. The job is the bottleneck.
September 2030.