The Consolation of Difficulty
A late friend had two sayings. The first one gets you started. The second one keeps you going. In the age of AI, the second one has gotten more useful, not less.
By Geordie Everitt
The first of Doug's two sayings was for the beginning: "Let's do something, even if it's wrong." I've written about him before — he died in a car accident in 1986, and I have been repeating both his sayings ever since. The first one I reach for when I need to break a paralysis. The second is for something different.
"If it was easy, everybody would do it."
These two cover different points in the arc of a task. The first gets you off the starting line. The second does its work later — after the novelty has worn off, after the difficulty has declared itself in full.
You know the moment. You are in the middle of something — a project, a skill, a commitment that is playing out across time — and the hard part is no longer speculative. You have been inside it long enough to know exactly what it requires. The gap between where you are and where you need to be has grown from estimate to fact. This is the stage where most people quit. Not from lack of ability. From a specific kind of demoralization: the discovery that the hard part is actually hard.
The Filtering Function
What Doug's second saying does is reframe the difficulty as information rather than indictment. Not evidence that you chose wrong. Not evidence that you are not suited for this. Evidence that the thing has content — that it repays effort because it requires effort. The difficulty is the moat. The moat is there because crossing it is not free.
Most worthwhile things require a sustained commitment that most people, at some point, decide is not worth the trouble. This is not a moral judgment. It is a structural observation about scarcity: the hard things are rare precisely because they are hard. If mastering a language required the same effort as choosing a playlist, everyone would speak three. If building something genuinely useful required no patience, patience would not be a competitive advantage. The difficulty is load-bearing.
What AI Changes About This
The natural objection, living through the decade we are in, is what happens to Doug's saying when artificial intelligence starts making hard things easy.
Part of the answer is that it reveals which difficulty was friction and which was the thing itself. Writing a passable memo was friction. Writing something with a point of view was always the thing. Debugging a stack trace was friction. Understanding a system deeply enough not to introduce the bug in the first place was the thing. When AI absorbs the friction, what remains tends to be the difficulty that was always worth the trouble.
There is also a less comfortable part of the answer. Some of the difficulty that gave people their professional identity was, it turns out, mostly friction. The hard part they took pride in is automatable. This is genuinely disorienting, and I don't want to paper over it. The correct response to that situation is not a philosophical reframe about "the new hard things." It is Doug's first saying, applied to the question of what to do next.
Two Sayings, One Philosophy
I have come to think of these two sayings as a complete set. Between them, they address the two existential crisis points that attach to any difficult undertaking.
Before you start: do something, even if it's wrong. Don't wait for a certainty that isn't coming.
When you're stuck in the middle: if it was easy, everybody would do it. The difficulty is not evidence of failure. It is evidence that you're doing something worth the trouble.
Doug apparently never said anything about what happens at the end. I've never needed to ask why. If you solve the beginning and the middle, the end tends to take care of itself.