Some additions to Google's robots.txt file:

------------------------
Disallow: /hosted/images/
Disallow: /hosted/life/
User-agent: Googlebot-Image
Disallow: /
User-agent: Googlebot-Image
Allow: /hosted/images/
Allow: /hosted/life/
Allow: /searchhistory/
Allow: /news?output=xhtml
Allow: /news?btcid=
Allow: /news?btaid=
Allow: /s2/profiles
------------------------

Why would they disallow the image search bot for most services? For example, this search will return 0 results:

http://images.google.com/images?q=site%3Agroups.google.com

... unless they'd "crawl" images internally and then add them to the index in a different way (considering they already have them on their servers, they don't strictly need to go the outside http route...). Not that that's the answer here, just saying...

If I see this right this part ...

   ---------
   User-agent: Googlebot-Image
   Disallow: /
   User-agent: Googlebot-Image
   ---------

has now been changed to ...

   ---------
   User-agent: Googlebot
   Disallow: /
   User-agent: Googlebot
   ---------

... and after yet another change those user-agent lines are now gone from the file completely.