diff --git a/tools for thinking.md b/tools for thinking.md
index 245811d37..7b010d5a8 100644
--- a/tools for thinking.md	
+++ b/tools for thinking.md	
@@ -107,17 +107,19 @@
         - you can get to 80% of the work fast this way
     - how does this compare to [[gpt-3]] generation? asker feels that AI based suggestions aren't really that great
       - can't speak about better/worse, but for some people this might feel easier
-    - [[core thesis for ai]]
-      - the AI can't be the product
-      - you won't print an essay ghost written by [[gpt-3]]
-        - -> yet? what about the recent paper that *was* written by [[gpt-3]], and for which they got first [[authorship]]?
-    - [[john borthwick]] many silos enforcing separation between ideas; also a risk of being fascinated by the technology or the tool
-      - [[jerry michalski]] is a good example of focusing on the thoughts, on the information, going beyond the [[tool]]
-      - it feels like the space has gone [[mainstream]] now
-        - but we're really still in the [[early adopter]] phase
-    - [[audio]]?
-      - grown a lot in audio
-      - 15-20% a year (check?)
+  - [[core thesis for ai]]
+    - the AI can't be the product
+    - you won't print an essay ghost written by [[gpt-3]]
+      - -> yet? what about the recent paper that *was* written by [[gpt-3]], and for which they got first [[authorship]]?
+  - [[john borthwick]] many silos enforcing separation between ideas; also a risk of being fascinated by the technology or the tool
+    - [[jerry michalski]] is a good example of focusing on the thoughts, on the information, going beyond the [[tool]]
+    - it feels like the space has gone [[mainstream]] now
+      - but we're really still in the [[early adopter]] phase
+  - [[audio]]?
+    - grown a lot in audio
+    - 15-20% a year (check?)
+    - transforming audio to a block or a highlight: there are early attempts, podcasting [[air]] or [[snipped]] (?)
+    - not a seamless experience
 
 ## [[Demo: Protocol Design for Tools for Thinking]]