diff --git a/TODO b/TODO
index 3a5617e..798243b 100644
--- a/TODO
+++ b/TODO
@@ -1,9 +1,12 @@
 ### GENERAL
+    * rm dups job should show progress bar
+    * in viewer, there is no move button (maybe add one?)
+    * consider doing duplicates before AI, and if there are say 100s+, then maybe pause the AI work
+        - had 5000+ new photos, took 8 hours to finish, for me to just delete them anyway
     * optimisation:
-        keep track of just new files since scan (even if we did this from the DB), 
+        - keep track of just new files since scan (even if we did this from the DB), 
         then we could just feed those eid's explicitly into a 'get_file_details_on_new_files'.
-        ALSO use new_eids list IF no new refimgs to do a  'run_ai_on_new_files' :)
-
+        - ALSO use new_eids list IF no new refimgs to do a  'run_ai_on_new_files' :)
 
     * allow changing dates in move dbox and then re-scan for existing folders OR just have a browse for existing...
         - for use on  scanned photos, they register as 2010, but are datestamped in visuals for 95
@@ -16,6 +19,18 @@
 
     * allow actions for wrong person:
         -> someone else? OR override no match for this person ever for this image?
+    * groups of persons (non-exclusive, recursive), so:
+        - fam (ddp, mandy, cam, mich)
+        - mandy_fam (mandy, kirsty, tony, anne, tenille)
+        - uncle_peters_fam (peter, joan, jarrod, aaron, alana)
+        - cousin_aarons_fam (aaron, karly, kai, etc)
+        - storviks (mandy_fam, uncle_peters_fam, cousin_aarons_fam, etc)
+
+    * birthdates in PA:
+        - M was interested in birthdate appearing in hass, so need an API /
+          sensor in hass too - maybe could store (optional) birthdate and
+          heads up (of X days, e.g. so can get present) or just come up on the day so you can remember to text them
+    * search logic (AND vs OR)
 
     * read this: https://flask.palletsprojects.com/en/2.2.x/testing/#faking-resources-and-context