Starting has never been cheaper
Most things don't fail. They get abandoned.
That was true before AI, and it's even more true now. Today anyone can spin up a prototype in an afternoon. You can ship a landing page by lunch and have a working demo by dinner. Starting has never been cheaper.
So starting isn't the advantage anymore. Finishing is.
Enthusiasm is common. Endurance is rare.
The last 20% is still yours
AI makes the first 80% feel easy. The last 20% is still yours: the edge cases, the users who don't get it, the bug that only shows up in production, the week where nobody signs up. That's where most AI projects quietly die. The tool didn't fail. The builder got bored.
Grit isn't intensity. It's not the all-night hackathon or the viral launch tweet. It's showing up on day 40, when the novelty is gone, the model keeps hallucinating, and nobody's watching.
It does not matter how slowly you go as long as you do not stop.
Stay with the problem, not the plan
Here's what changes with AI: persistence compounds faster. Every day you stay with a problem, your tools get better, your prompts get sharper, and your understanding goes deeper. One day of work looks like nothing. A year of it, with AI multiplying every hour, looks like luck to everyone who missed the boring middle.
But persistence isn't stubbornness. AI moves too fast for that. The model you built on six months ago may already be outdated. So keep the goal and change the method. Stay with the problem, not with the plan, the stack, or the prompt that already failed.
How do we move forward?
Pick a problem worth staying with.
AI lets you chase ten ideas at once. Don't. Choose one you'd still care about in two years.
Let AI handle the speed, and you handle the direction.
Tools can write code. They can't decide what's worth building.
Measure by what you've finished, not what you've started.
A shipped project beats ten demos.
Expect the boring middle.
It isn't a sign you're failing. It's the part everyone else skips.
Be the one who finishes
A simple test I use: would I still do this if it took twice as long? If yes, keep going. If no, it was never the thing.
AI will make the world full of smart people with unfinished work. Be the one who finishes.
Nothing in the world can take the place of persistence.