diff --git a/TODO.md b/TODO.md
index e3c24df..06dd1e3 100644
--- a/TODO.md
+++ b/TODO.md
@@ -1,6 +1,6 @@
 # system-a
 
-[system-a] rename `sqrt` to `squareRoot`
+[the-book] rename `sqrt` to `squareRoot`
 [system-a] `gradientDescentRms`
 
 # the-book
diff --git a/src/system-a/gradient-descent/gradientDescentRms.ts b/src/system-a/gradient-descent/gradientDescentRms.ts
index b593725..416bc45 100644
--- a/src/system-a/gradient-descent/gradientDescentRms.ts
+++ b/src/system-a/gradient-descent/gradientDescentRms.ts
@@ -1,5 +1,5 @@
 import { tensorZeros, type Tensor } from "../tensor/index.js"
-import { add, div, mul, sqrt, square, sub } from "../toys/index.js"
+import { add, div, mul, square, squareRoot, sub } from "../toys/index.js"
 import type { Representation } from "./Representation.js"
 import { gradientDescent } from "./gradientDescent.js"
 import { smooth } from "./smooth.js"
@@ -19,7 +19,7 @@ export function gradientDescentRms(options: {
       // NOTE Add `stabilizer` to avoid `div` by zero.
       const adaptiveLearningRate = div(
         options.learningRate,
-        add(stabilizer, sqrt(r)),
+        add(stabilizer, squareRoot(r)),
       )
       return [sub(p, mul(adaptiveLearningRate, g)), r]
     },
diff --git a/src/system-a/toys/toys.ts b/src/system-a/toys/toys.ts
index e418399..2e91f20 100644
--- a/src/system-a/toys/toys.ts
+++ b/src/system-a/toys/toys.ts
@@ -34,7 +34,7 @@ export const exptScalar = prim2(
   (ra, rb, z) => [rb * ra ** (rb - 1) * z, ra ** rb * Math.log(ra) * z],
 )
 
-export const sqrtScalar = prim1(
+export const squareRootScalar = prim1(
   (x) => Math.sqrt(x),
   (ra, z) => (1 / 2) * ra ** (-1 / 2) * z,
 )
@@ -51,7 +51,7 @@ export const mul = extend2(mulScalar)
 export const div = extend2(divScalar)
 export const log = extend1(logScalar)
 export const expt = extend2(exptScalar)
-export const sqrt = extend1(sqrtScalar)
+export const squareRoot = extend1(squareRootScalar)
 export const square = extend1(squareScalar)
 
 export const lt = comparator((x, y) => x < y)