Table of contents
Open Table of contents
The best kind of lazy
There’s a type of lazy I deeply respect. Not the “avoid work” lazy — the “why are we doing this at all” lazy. The person who looks at a manual process and says, “this is a waste of time” and then spends an afternoon making it go away forever.
I have a friend who automated his entire monthly reporting workflow in a weekend. It used to take him six hours on the first Monday of every month — pulling data from four systems, formatting charts, writing commentary. Now a script runs on the first of the month and drops a finished report in his manager’s inbox. Six hours per month, seventy-two hours per year, reclaimed permanently.
He didn’t do it because he’s a hard worker. He did it because he’s lazy — strategically lazy. He recognized that manual repetition is a design failure, not a job requirement, and he fixed the design.
AI made lazy accessible
Before Claude, being strategically lazy required a specific kind of effort. You had to learn a scripting language, or figure out cron jobs, or wrestle with API documentation for an afternoon. The activation energy for automation was high enough that most people just accepted the manual work.
Now? Describe the boring thing you do every week, ask Claude to write a script for it, spend ten minutes testing, and you’re done. The activation energy collapsed.
In one recent month I automated three things I’d been doing by hand for years:
- A Slack reminder system for stale PRs (manual nagging → automated)
- A release-notes generator that drafts the changelog from merged PRs (an hour each release → a glance)
- A local dev environment setup script (thirty-minute onboarding → one command)
Total time invested: maybe four hours. Total time saved per month: roughly eight hours. That’s a 24x annual return on a one-time investment. Lazy math is good math.
Before AI, being strategically lazy required specific technical effort. Now it requires noticing the repetition and describing it.
The principle: never do the same thing manually three times
I have a simple rule. If I catch myself doing something manually for the third time, I stop and automate it. Not the tenth time. Not “when I have time.” The third time. Because by the third repetition, I know the pattern. I know the inputs and outputs. I know what a script needs to do.
Claude makes this rule practical. What used to be a “I’ll automate this when I have a free weekend” becomes a “give me five minutes and a prompt.” The ROI on five minutes of prompt-writing versus hundreds of future manual repetitions is infinite.
Here’s the trick: you have to notice the repetition. Most of us are so deep in our habits that we don’t register when we’re doing the same thing for the tenth time. Start paying attention. Every time you think “ugh, not this again” — that’s a signal. That’s your cue to automate.
What strategic laziness is not
Let me be clear about what I’m not saying. I’m not saying “do less work.” I’m not saying “cut corners.” I’m not saying “let AI do your job for you and collect a paycheck.”
Strategic laziness is redirecting effort from low-leverage to high-leverage work. It’s recognizing that manually formatting a report is low-leverage — a machine can do it. Designing the report’s structure, choosing what to measure, deciding what the numbers mean — that’s high-leverage. That’s where your brain should be.
The people I know who practice this aren’t doing less. They’re doing different. They’ve shifted their time from execution to strategy, from repetition to creation, from maintenance to innovation. The total effort might be the same. The output is dramatically higher.
Strategic laziness isn’t doing less. It’s redirecting effort from what machines handle well to what only humans handle well.
How to start being strategically lazy
If any of this resonates, here’s a concrete approach:
Audit your week. For one week, keep a simple log. Every time you do something that feels repetitive, write it down. Not formal. Just a bullet in a note. By Friday, you’ll have a list of automation candidates.
Rank by frequency and frustration. Some tasks happen daily and feel awful (high priority). Some happen monthly and are mildly annoying (low priority). Start with the daily awful ones — they compound fastest.
Describe, don’t script. Don’t try to write the automation yourself. Describe the task to Claude in plain language: “Every Monday I do X by opening Y, copying Z, and pasting it into W. Write a script that does this.” Let the tool figure out the implementation.
Test once, trust with caution. Run the script once, verify the output manually, then let it run on schedule. But check the output periodically — automation that silently produces wrong results is worse than no automation.
The meta-lesson
There’s a deeper point here that applies beyond automation. The skill of the AI era isn’t doing more things — it’s identifying which things are worth doing and which things are worth delegating. The delegation used to require a teammate. Now it just requires a clear description.
The engineers who thrive won’t be the ones who work the hardest. They’ll be the ones who are smartest about where they direct their effort. They’ll automate the repetitive, delegate the mechanical, and reserve their own attention for the decisions that actually shape the product. That redirection — effort flowing from what machines handle well to what only judgment handles — is the same bet behind my whole AI Engineering practice.
Call it lazy. I call it leverage. And the tools have never been better.