Fix typo and add 'experimental' clarification. (tensorflow#149)

docs/DifferentiableFunctions.md

@@ -4,6 +4,15 @@
 
 Last updated: March 2019
 
+> #### Experimental
+>
+> Automatic differentiation and differentiable programming are being incubated in the
+> ['tensorflow' branch of apple/swift](https://github.com/apple/swift/tree/tensorflow)
+> and released as part of the
+> [Swift for TensorFlow toolchains](https://github.com/tensorflow/swift#getting-started),
+> which you can play with. The authors will propose this feature through
+> [Swift Evolution](https://forums.swift.org/c/evolution) in 2019.
+
 ## Introduction
 
 Swift supports differentiable functions as part of the language. The
@@ -24,9 +33,9 @@ can be differentiated.
 APIs that take `@differentiable` functions, like [`gradient(of:)`]. The binary
 representation of a `@differentiable` function is a special data structure
 containing the original function along with extra information required for
-computing its derivatives. Usage `@differentiable` functions are a part of
-Swift's type system. Most notably, they are used by differentiation APIs in the
-standard library. Here are some examples demonstrating differentiation APIs:
+computing its derivatives. `@differentiable` functions are a part of Swift's
+type system. Most notably, they are used by differentiation APIs in the standard
+library. Here are some examples demonstrating differentiation APIs:
 
 ```swift
 func square(_ x: Float) -> Float {

docs/DifferentiableTypes.md

@@ -4,6 +4,15 @@
 
 Last updated: March 2019
 
+> #### Experimental
+>
+> Automatic differentiation and differentiable programming are being incubated in the
+> ['tensorflow' branch of apple/swift](https://github.com/apple/swift/tree/tensorflow)
+> and released as part of the
+> [Swift for TensorFlow toolchains](https://github.com/tensorflow/swift#getting-started),
+> which you can play with. The authors will propose this feature through
+> [Swift Evolution](https://forums.swift.org/c/evolution) in 2019.
+
 ## Preface
 
 Speaking in terms of elementary calculus, only functions are "differentiable": only functions have derivatives and can be differentiated. In this document, the terminology "differentiable types" is used as a shorthand for "types that can be used as arguments and results of differentiable functions". This notion is important because not all types are "differentiable" in this sense. For example, types representing real numbers and vector spaces are "differentiable", but strings and integers are not.

docs/DynamicPropertyIteration.md

@@ -4,6 +4,16 @@
 
 Last updated: March 2019
 
+> #### Experimental
+>
+> `KeyPathIterable` is being incubated in the
+> ['tensorflow' branch of apple/swift](https://github.com/apple/swift/tree/tensorflow)
+> and released as part of the
+> [Swift for TensorFlow toolchains](https://github.com/tensorflow/swift#getting-started),
+> which you can play with. The authors will propose this feature through
+> [Swift Evolution](https://forums.swift.org/c/evolution) in 2019. Updates will be posted
+> on the [intial pitch thread](https://forums.swift.org/t/storedpropertyiterable/19218).
+
 ## Background and motivation
 
 The ability to iterate over the properties of a type is a powerful reflection