diff --git a/talk/popl-winter-2019/outline.org b/talk/popl-winter-2019/outline.org
index e01c3ae51c2c9bf9dbe3cc8af278ea783d530fb4..10651dc4779d78e4b312551ed63b7b4f967d1eff 100644
--- a/talk/popl-winter-2019/outline.org
+++ b/talk/popl-winter-2019/outline.org
@@ -133,6 +133,11 @@ Next: Axioms
       - emphasize the axioms to derivations more
       - emphasize reasoning principles rather than type safety
     - future work: more complex type features
+*** Problems
+    - Weak transition in intro
+      - From there are a bunch of semantics to our contributions
+    - Inconsistent terminology
+      - soundness principles, type soundness, semantic principles
 ** Questions
    - What about blame?
      - We don't give a treatment of blame in the paper, but a paper by
@@ -141,7 +146,7 @@ Next: Axioms
        positive party
      - And we argue in the paper that that is related to our proof
        that upcasts are pure and downcasts are linear
-   - Even CBV Eta is violated by pointer/object equality?
+   - Eta is violated by pointer/object equality?
      - right, and for this reason the simple wrapping semantics of
        contracts/casts doesn't work for those systems, and people
        introduce proxies
diff --git a/talk/popl-winter-2019/popl-talk.key b/talk/popl-winter-2019/popl-talk.key
index adcdc9bb0b66b7fc2079b0bd89800a015461d993..594bcb9c60091926250c335a3536c70e62893c0a 100644
Binary files a/talk/popl-winter-2019/popl-talk.key and b/talk/popl-winter-2019/popl-talk.key differ