diff --git a/2018/Microsoft.ML.Core/README.md b/2018/Microsoft.ML.Core/Session1.md
similarity index 98%
rename from 2018/Microsoft.ML.Core/README.md
rename to 2018/Microsoft.ML.Core/Session1.md
index 805d95d..d4e8e10 100644
--- a/2018/Microsoft.ML.Core/README.md
+++ b/2018/Microsoft.ML.Core/Session1.md
@@ -102,6 +102,7 @@ ML.NET.
     - Weekly review (Thursdays)
     - 2 x four hours with lunch in between
 * Review other core assemblies:
-    - `Microsoft.ML.CpuMath`
+    - `Microsoft.ML.Core`
     - `Microsoft.ML.Data`
-    - `Microsoft.ML.Transforms`
\ No newline at end of file
+    - `Microsoft.ML.Transforms`
+    - `Microsoft.ML.CpuMath`
\ No newline at end of file
diff --git a/2018/Microsoft.ML.Core/Session2.md b/2018/Microsoft.ML.Core/Session2.md
new file mode 100644
index 0000000..d4ac5d9
--- /dev/null
+++ b/2018/Microsoft.ML.Core/Session2.md
@@ -0,0 +1,69 @@
+# Microsoft.ML.Core
+
+Status: **Needs more work** | 
+[API Ref](Microsoft.ML.Core.md) |
+[Dependencies](Microsoft.ML_Dependencies.png) |
+[Write-up](https://github.com/TomFinley/machinelearning/blob/a4511133d3c0b5bc993af22f09607945e7fdf063/docs/code/Catalog-Core.md)
+
+## Notes
+
+* `Column` can be a struct
+    - Sharing is no longer a concern
+* `Schema`
+    - Should probably have a struct based enumerator
+    - The value tuples don't play nicely with F#
+    - We should probably remove them
+* `Metadata`
+    - Key/Value pairs associated with a single column
+    - Uses `Schema` to describe itself, which is powerful, but might make it
+      hard for developers to reason about. Can we simplify this? It seems the
+      answer is no (concerns around ownership as well as benefits that are
+      gained by uniformity with other parts, such as be able to persist the
+      schema).
+    - *Metadata* is a fairly generic term; the entire schema is metadata. Can we
+      come up with a more specific name? Like annotations?
+    - The getter seems complicated but it allows them to not keep the metadata
+      instances around and they allow for easier chaining.
+    - We could have a helper method or property that gets the value in an easier
+      way, but it's hard to make the much simpler as they would have to be
+      generic or the user would have to handle the type
+* `IDataView`
+    - `ISchematized` is being removed
+    - Should be an abstract class
+    - `GetRowCount()`. `lazy = true` means get if you know it, otherwise don't
+      bother. `lazy = false` means, the caller *really* would like to know, but
+      implementers can still return `null`.
+        - Either make it a property `long? Count` or make it
+          `bool TryGetCount(out long count)`.
+    - `GetRowCursor()` instead of taking `Func<int, bool>` make it `Predicate<Column>`
+      which makes it easier to reason about.
+    - We should look into whether we should have an async version
+    - We want to get rid of the the `IRowCursorConsolidator` base
+* `ICounted`
+    - It seems it's not taken by anyone
+    - It seems it's co-implemented by `IRow` and `ICursor`
+    - Can this be collapsed?
+* `IRowCursor` is needed, but can we get rid of `ICursor`?
+* `UInt128` should be something more specific like `RowId`
+* `GetRootCursor()` is an performance optimization because many cursors just
+  delegate. This could also be solved by `GetMover()`. If we converted `ICursor`
+  to an abstract class, we might be able to hide this in the implementation, or
+  at least make it protected.
+* `MoveMany()` seems like something that could be removed.
+* `ICursor.State` seems to be unused, can this be removed?
+* `IRow.IsColumnActive` seems to be mostly used in asserts, is this needed? It
+  seems `GetGetter()` could throw or return `null`.
+* `IRowCursor` extending `IRow` seems odd. Can these be collapsed? If not, can
+  they have an inheritance relationship when we're making them classes?
+* `IRowToRowMapper`
+    - `Action disposer` should probably be replaced by making the right type
+      (`IRow`?) implement `IDisposable`.
+    - `Func<int, bool>` is hard to get right. Should this be `Predicate<Column>`
+      as well?
+* Can we remove `DataKind` entirely? If not, we should probably rename some of the
+  members, remove the overlap, and add an entry for `0`, like `Unknown` or `None`.
+* `ColumnType`
+    - We should remove the many `IsXxx` overloads, especially for the esoteric
+      types.
+    - It seems this to be heavy context; is it possible to reduce the number
+      moving pieces, such as removing the concrete derivatives?
\ No newline at end of file