For the first couple of years of my career, I wore long hours like a badge. I was usually the last one online, I answered messages at midnight, and I genuinely believed that the amount of time I put in was the measure of how good I was. I was tired a lot, and I was proud of being tired. It took me an embarrassingly long time to notice that the engineers I admired most weren't doing this — and were getting more done.
This is about the shift from measuring effort in hours to measuring it in outcomes, and why smart work beats long hours almost every time in software.
The story
The moment it clicked was uncomfortable. A senior engineer on my team consistently delivered more impact than I did, and he logged off at a normal hour. I'd quietly assumed he was just more talented. Then I actually watched how he worked, and it wasn't talent — it was choices.
He spent the first part of his day deciding what mattered, and then protected long blocks to do that one thing well. I spent my day reacting — to messages, to whatever was loudest, to a dozen half-tasks. He automated things he did more than twice. I did them manually, every time, because setting up the automation felt like a detour. He said no to work that didn't matter. I said yes to everything to look helpful, then worked late to survive the pile I'd accepted.
He wasn't working harder than me. He was working on the right things, with fewer interruptions, after eliminating repetitive work. My long hours weren't a sign of dedication — they were a symptom of bad prioritization that I was papering over with time. The late nights were the cost of being busy instead of effective.
Why long hours fail in engineering
Software isn't a job where output scales linearly with hours. It's creative, judgment-heavy work, and the quality of decisions matters far more than the quantity of time. One good architectural decision can save weeks. One tired, sloppy decision at 1am can cost them. I shipped some of my worst bugs late at night, "pushing through," and then spent fresh-morning hours cleaning them up. The long hours were literally creating the work that justified more long hours.
In creative work, exhausted time is low-quality time. Ten focused hours beat sixteen tired ones, and they don't cost you the next day too.
There's also a quieter problem: when you glorify long hours, you optimize for looking busy instead of being effective. You take the visible, reactive work because it feels productive, and you avoid the invisible, high-leverage work — thinking, automating, simplifying — because it doesn't look like "working." That's exactly backwards.
What smart work actually looks like
Prioritize ruthlessly. Decide what actually matters before you touch the keyboard. Most of the impact comes from a small fraction of the work; long hours usually mean you're spending time on the rest.
Protect deep focus. A few uninterrupted hours produce more than a whole fragmented day. Guard them like they're the job, because they are.
Automate anything you do repeatedly. The detour to automate pays back fast. I used to manually do things I'd done a hundred times; now if I do something twice, I ask whether the third time should be a script.
Use leverage. AI for the boring parts, good tooling, reusable foundations, asking the person who already knows instead of struggling for an hour. Leverage is how you get more output from less time without working harder.
Say no. Every yes to unimportant work is a no to important work — usually paid back with your evening. Protecting your time is protecting your impact.
Key lessons
Hours are an input; outcomes are the point. Nobody who matters remembers how long it took. They remember what shipped and whether it worked.
Tired time is expensive time. Late-night pushing often creates more work than it finishes.
Busy is not effective. The most visible work is rarely the most valuable. Optimize for impact, not for looking occupied.
Leverage beats effort. Automation, tooling, focus, and good prioritization compound. Raw hours don't.
Action steps
- Start each day by choosing one or two things that actually matter. Do those first, before the noise.
- Block focus time and defend it. No meetings, no messages, for a real stretch. Treat it as non-negotiable.
- Automate the third repetition. If you've done something twice, the next time should be scripted or templated.
- Practice saying no (or "not now"). Protect your capacity for the work that counts.
- Judge your day by what moved, not by hours logged. If you shipped something that mattered in six focused hours, that's a great day — log off.
Final thoughts
I'm not against hard work — I work hard. But I stopped confusing hard work with long hours. The goal was never to be tired; it was to make an impact, and the two are often opposites. The engineers and teams that win sustainably are the ones who protect their focus, eliminate the repetitive, say no to the unimportant, and spend their best energy on the work that actually matters.
Working late occasionally for something that counts is fine. Living there is just a slow way to do less while feeling like you're doing more. Work smart, protect your focus, and go home.