Grit in a World of AI

AI is astonishingly good at the first 80% of almost anything now. The draft. The outline. The plausible-sounding plan. The version that’s fine.

Which means the whole game of building something has quietly moved to the other 20% — the part that was always the hard part anyway.

That part still runs on the least scalable resource there is: your willingness to keep going when it stops being fun.

Angela Duckworth, who spent a career studying this, defines grit as passion plus perseverance toward a long-term goal. Her recent point — that grit matters more in an AI world, not less — landed with me, because it names something founders are feeling but haven’t quite said out loud: the easy parts got easier, and it made the hard parts feel harder by contrast.

What AI won’t do for you

Here’s what I mean. AI will write you a decent first draft of a launch email in nine seconds. It will not sit with the discomfort of realizing, on draft four, that the whole positioning is wrong and you have to start over. It’ll generate forty campaign ideas. It won’t make the thirty cold follow-ups after thirty-eight of them go nowhere. It’ll produce the plan. It won’t do the plan on a Tuesday when the plan is boring and no one is watching.

The middle mile

The un-automatable middle mile — the tedious, unglamorous, no-dopamine stretch between “great idea” and “actually shipped and working” — is exactly the stretch AI can’t walk for you. It just made the trailhead so smooth that more people start, which means the finish line is more crowded with abandoned first drafts than ever.

I find this weirdly encouraging. For years the advantage went to whoever could produce the most, fastest. That race is now essentially over — a machine won it. What’s left is a much older, more human advantage: caring about a specific thing enough to stay with it after the novelty burns off.

Friction was doing quiet work

There’s a small irony in here for anyone building a lean company. We adopt AI to remove friction — and friction, it turns out, was doing some quiet work. The struggle to make a thing was part of how we figured out whether we believed in it. When the struggle to produce disappears, you have to go find your conviction somewhere else, on purpose. The grit isn’t in the making anymore. It’s in the deciding, and the staying.

So if you’re a founder feeling faintly guilty that AI does half your job now and it still feels hard — that’s not a you problem. The job was never the typing. The job was the judgment and the perseverance, and those were always the expensive parts. AI just stripped away the busywork that used to disguise how much of the work was really that.

The machines will keep getting better at the 80%. Good. Let them. It frees you up for the 20% that was always going to decide whether this thing works — and that part, thankfully, is still yours.