diff --git a/demo.ipynb b/demo.ipynb
index bf0ebd5..80a656a 100644
--- a/demo.ipynb
+++ b/demo.ipynb
@@ -235,7 +235,7 @@
     "\n",
     "This notebook demonstrates how to make a differentially private release with [OpenDP](https://docs.opendp.org) using the data you've provided. The OpenDP programming interface has many more options than we can expose in a web interface, and this notebook will suggest some options to explore.\n",
     "\n",
-    "Differential privacy quantifies how much privacy is lost when answering/releasing queries about a sensitive dataset. Quoting Dwork and Roth, \"giving overly accurate answers to too many questions will inevitably destroy privacy.\" With this in mind, it is customary to set what is called a \"privacy budget\" that limits the number and accuracy of queries. In practice, this means you need to be careful how you spend your privacy budget, because you won't be able to release queries that will cause you to exceed your privacy budget.\n"
+    "Differential privacy quantifies how much privacy is lost when answering/releasing queries about a sensitive dataset. Quoting Dwork and Roth, \"giving overly accurate answers to too many questions will inevitably destroy privacy.\" With this in mind, it is customary to set what is called a \"privacy budget\" that limits the number and accuracy of queries. In practice, this means you need to be careful how you spend your privacy budget, because you won't be able to release queries that will cause you to exceed your privacy budget.\n",
     "Interactively developing DP analyses is a good practice, for example, to identify variables of interest before choosing where to spend the rest of a finite privacy budget. That said, it is important to track the total budget consumed: It would not be good to run a cell in this notebook, look at output, and then run it again, without tracking the privacy consumed by each operation.\n",
     "\n",
     "We begin with a conventional set of imports:"