diff --git a/python/least_squares.ipynb b/python/least_squares.ipynb
index 256ecbde..e382cdc7 100644
--- a/python/least_squares.ipynb
+++ b/python/least_squares.ipynb
@@ -234,10 +234,10 @@
     "$$\n",
     "\\begin{aligned}\n",
     "x^* &= \\arg \\min_x f(x)\\\\\n",
-    "\\textrm{s.t.} &\\quad l \\succcurlyeq x \\succcurlyeq u\n",
+    "\\textrm{s.t.} &\\quad l \\preccurlyeq x \\preccurlyeq u\n",
     "\\end{aligned}\n",
     "$$\n",
-    "Where $l$ and $u$ are respectively lower and upper bound vectors and the inequalities $l \\succcurlyeq x \\succcurlyeq u$ are read elementwise. Solving for the minimizing $\\delta x$ is now a bit more involved than solving a linear system, and the optimization method is now referred to as [Sequential Quadratic Programming](https://en.wikipedia.org/wiki/Sequential_quadratic_programming) (SQP). Fortunately MuJoCo has an efficient box-constrained QP solver: the [mju_boxQP](https://mujoco.readthedocs.io/en/latest/APIreference/APIfunctions.html#mju-boxqp) function."
+    "Where $l$ and $u$ are respectively lower and upper bound vectors and the inequalities $l \\preccurlyeq x \\preccurlyeq u$ are read elementwise. Solving for the minimizing $\\delta x$ is now a bit more involved than solving a linear system, and the optimization method is now referred to as [Sequential Quadratic Programming](https://en.wikipedia.org/wiki/Sequential_quadratic_programming) (SQP). Fortunately MuJoCo has an efficient box-constrained QP solver: the [mju_boxQP](https://mujoco.readthedocs.io/en/latest/APIreference/APIfunctions.html#mju-boxqp) function."
    ]
   },
   {