Interesting concept. If you don't mind, some comments on your site/service:
1. The slides on the main page aren't showing up for me in FF or Chrome.
2. You're targeting avid readers, and are trying to save them time, but you're pushing video how-to's? Text + screen shots, please! (Or maybe A/B testing will prove me wrong)
3. Why are all the images of articles tilted and in such small font? I want to see how well your service works for me, and what's better than reading an article it's analyzed?
4. Your how-to for RSS scares me a bit. It goes by so fast, and I have to copy-paste URLs? I have to find RSS URLs that as this article mentions, are getting increasingly hard to find? I'm already subscribed via Google Reader. I could copy the URL from the 'Details and Statistics' of a feed, but why can't you use their API to bulk import my feeds (competitors to Reader do this). I'm not sure if you could automagically rewrite Reader subscription URLs once scraped, but it seems like the easiest option to me (as long as you have an intermediate page where I can exempt certain feeds.
5. Is there any way I can see the results of your bookmarklet on a few actual articles before signing up? Even if it's just a mirror of some popular articles with the highlighting precomputed, as long as it's the whole article. I feel like you have a lot of work to do to convince someone that they're not going to miss some crucial bit of information by ignoring what you don't highlight.
7. Why do I need to remember another username? Is this a 'social' service? You already have my email!
8. It doesn't seem to work with Readability
9. Why do you need to track me with cookies? You already have me running javascript on every page I use the bookmarklet on...but I appreciate being told about them
10. How are you going to make money?
(just used the bookmarklet on this Wired article[1] currently on the front page of HN. Very surreal. And I'm not sure if I got the main point of the article or not...)
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
3. It might be good to keep the tilted pics in the slideshow, and straight pictures (or actual text) elsewhere. Or maybe tilted everywhere! I can tell you what bothers me, but what really matters is how many clickthroughs you get, right?
7. IMO, it shouldn't be social (how would it be?) so no usernames are necessary, but I'm probably not your average user. Nor have I ever created an online service/business, so keep that in mind when following my advice.
9. I don't know about bothered, since I have cookies enabled for everything else. It'd be less weird if you made it clear the cookie didn't track me across other sites,
10. No, just want to make sure you're thinking about it before possibly getting attached to a service that might not survive (e.g. you get bought and the buyer shuts it down). If I found the service useful (and especially if I were a CEO/executive, I think), I would love to pay for it, either a 1-time fee (like I did with Pinboard.in), or yearly subscription. Crazy idea: maybe you could license the software to news sites so they can have built-in highlighting with the click of a link (similar to 'print-friendly' links every article has)
My RSS reader has few enough feeds that I've been able to read them all, but I've been careful to only add feeds that I could keep up with (with keenskim, that might change). Other than that, I stick to the top of HN.
1. The slides on the main page aren't showing up for me in FF or Chrome.
2. You're targeting avid readers, and are trying to save them time, but you're pushing video how-to's? Text + screen shots, please! (Or maybe A/B testing will prove me wrong)
3. Why are all the images of articles tilted and in such small font? I want to see how well your service works for me, and what's better than reading an article it's analyzed?
4. Your how-to for RSS scares me a bit. It goes by so fast, and I have to copy-paste URLs? I have to find RSS URLs that as this article mentions, are getting increasingly hard to find? I'm already subscribed via Google Reader. I could copy the URL from the 'Details and Statistics' of a feed, but why can't you use their API to bulk import my feeds (competitors to Reader do this). I'm not sure if you could automagically rewrite Reader subscription URLs once scraped, but it seems like the easiest option to me (as long as you have an intermediate page where I can exempt certain feeds.
5. Is there any way I can see the results of your bookmarklet on a few actual articles before signing up? Even if it's just a mirror of some popular articles with the highlighting precomputed, as long as it's the whole article. I feel like you have a lot of work to do to convince someone that they're not going to miss some crucial bit of information by ignoring what you don't highlight.
7. Why do I need to remember another username? Is this a 'social' service? You already have my email!
8. It doesn't seem to work with Readability
9. Why do you need to track me with cookies? You already have me running javascript on every page I use the bookmarklet on...but I appreciate being told about them
10. How are you going to make money?
(just used the bookmarklet on this Wired article[1] currently on the front page of HN. Very surreal. And I'm not sure if I got the main point of the article or not...)
[1] http://m.wired.com/gadgetlab/2012/04/can-an-algorithm-write-...