Now for the answers:
1. my bad. I was doing some optimizations on the web server (via http://gtmetrix.com ). Now it should work
2. true! received same feedback yesterday. will change.
3. just thought it would look better. but that doesn't mean better experience. will change them (for bigger images just click them)
4. originally KeenSkim was built for RSS feeds. After starting the alpha testing we pivoted to the webpage bookmarklet. It works so well that we didn't improve the RSS part. To be fixed in the next weeks (opml support)
5. Not yet but great idea! we'll do that
7. It's not social (not built with this intention in mind, but could be). Social auth and simpler login are in the top 3 priorities after stabilizing the summarization algorithm
8. will look on that
9. the cookies are for the site, not the bookmarklet (cross-domain calls). It's mainly for login (standard Django). Does it bother you in any specific way?
10. mainly a freemium approach. the exact business model is not clear (we'll test it after we get an idea about usage, costs, growth, ...). Do you have anything in mind?
About the results: there are pages on which it work good, on some pages great and on some really bad. The problem on this article is that it has several story-lines and KeenSkim doesn't work well on that (interviews, lists, tutorials won't be properly summarized)
EDIT: There is still space for improving the service. That's priority no. 1
Where there other news that you read?
Thanks again!