Update HowItWorks.md

docs/HowItWorks.md

@@ -306,7 +306,7 @@ Note that for this error case, it is more of an art than a science. That is to s
 would add special casing for anticipated scenarios, rather than construct some general-purpose rule.
 
 #### Other notes
-* Just as it's imporant to understand the assumptions that *will* hold true,
+* Just as it's important to understand the assumptions that *will* hold true,
 it is also important to understand the assumptions that will not hold true.
 One such **non-invariant** is that not every token generated by the lexer ends up in the tree.
 
@@ -387,6 +387,6 @@ existing, more battle-tested, work to validate our own correctness.
 degree of variance, we should set up some infrastructure to help us ensure that performance work results in a statistically
 significant boost and works on a wide variety of machine configurations.
 * Fuzz testing - test the parser against automatically generated inputs to exercise edge cases in a bulk fashion, and
-and ensure expected properties of the tree.
+ensure expected properties of the tree.
 * Community feedback - try and get the parser in the hands of as many people as possible so we can validate a
 wide range of use cases. The Syntax Visualizer tool is one tool to help us increase reach.