Poke the Box by Seth Godin — A Review

Seth Godin published Poke the Box in 2011, when the barrier to shipping something — anything — was still significant. You needed capital, distribution, manufacturing, an audience you'd spent years building. The costs of starting were high enough that inaction had a rational justification.

That world is gone. The friction to ship has dropped to nearly zero. Want to start a newsletter? It takes twenty minutes. Want to test a product idea? You can have a landing page, a payment link, and a basic AI-assisted prototype running before the end of the day. Want to launch a podcast? The total investment is a decent microphone and a free account on a hosting platform.

And yet. Look around at the people who should be starting things and aren't.

The Real Barrier Was Never the Tools

Godin's argument in Poke the Box is that the barrier to starting was never primarily external. It was always the fear of being wrong, being judged, being the person who tried something and failed. The external friction was real — but it also served as a convenient excuse that let people avoid confronting the internal resistance.

Now the external friction is mostly gone. Which means the internal resistance is the only thing left, and it's completely visible. The person who says 'I'd start something if only I had the time / the money / the distribution / the platform' — in 2026, those excuses don't hold.

The cost of being wrong is less than the cost of doing nothing.

What I Took From This

The specific permission I got from this book was to treat incomplete, imperfect starts as the correct unit of progress. Not the well-planned launch. Not the polished product. The deliberate, visible, iterative experiment.

I've started more things badly since reading this than I did before, and the success rate on things I started badly and then improved is dramatically higher than the success rate on things I waited until I was confident about.

The AI Irony

There's a specific irony I think about with this book in the AI era. The tools have never been more capable. The cost of experimenting has never been lower. Building a minimum viable version of almost anything — a product, a service, a piece of content — can be done in hours with the right AI assistance. Which means the only remaining excuse for not starting is the one that was always the real excuse anyway: fear.

Read this when you're stuck on something. It's short enough to finish in an afternoon, and it will annoy you in exactly the right way.