---
title: "Momentum opens doors"
description: "Compounding momentum eventually crosses every finite threshold, a useful way to think about AI and startups."
canonical: "https://www.rickblalock.dev/scratch-notes/momentum-opens-doors/"
date: "2026-09-23"
updated: "2026-09-23"
author: "Rick Blalock"
---

# Momentum opens doors

> Compounding momentum eventually crosses every finite threshold, a useful way to think about AI and startups.

Published: 2026-09-23

Catalog: SN-014

[DHH wrote](https://x.com/dhh/status/2099481279045697850): “Momentum opens all doors.”

Is there a formula for that? No idea, but I played around with it a bit in my head.  The best proof I can provide, I'll sum up this way:

> Compounding momentum eventually crosses every finite threshold.

Start with positive momentum and let it compound at a steady positive rate:

$$
M_t = M_0(1+r)^t
$$

$M_0$ is your starting momentum, $r$ is the growth rate, and $t$ is time. A door opens when your momentum reaches its barrier, $D$:

$$
M_t \ge D
$$

Keep that compounding going, and any fixed, finite barrier eventually gets crossed. You don’t have to start above the threshold to reach it.  (Caveat for the nerds: Obviously there's always some unseen barrier in physics..but you get the point).

Regarding AI, startups, and this new crazy world we live in: A capability can improve for a while before it becomes reliable or affordable enough to build a business around. Something that was too expensive to attempt becomes a reasonable thing for a small team to try.  And that's exactly what this new AI stuff gives us.  Even services now has a completely new set of scaling laws, which I don't think anyone has fully proven out yet.

Anyway back to momentum: A team can build momentum of its own. Getting a working product in front of a customer sooner means learning sooner. Feedback improves the next version. A customer who gets value can become a reference for another. Each win can make the next one easier, while improvements in AI expand what the team can take on.  This is all well known and written about but with AI, the speed is so fast that you end up just going around barriers sometimes and not even consider them anymore.

The math assumes the compounding continues and the barriers stay fixed. A startup still has to find demand and survive long enough to benefit. But today’s limitations aren’t permanent. There’s a reason to keep working on a problem you understand, even when some of what you want to do remains out of reach.

---

[All Scratch Notes](https://www.rickblalock.dev/scratch-notes/)