Cool project.

For the "tutorial", I read through maybe the first 3 boxes, then got tired and clicked until I could explore it myself. I bet it's the same for most users (feel free to measure to get real data). If you can, defer teaching to mouseovers and other hints in the UI.

This is very similar to something I've been looking for/thinking of building for some time. I took a look around but browsing was very slow for me so I'll just ask here; do you have any plans for training filters? That is, feeding it data about what sort of thing you read (and on what basis; software announcements I might click on daily, but if I haven't checked in for a month I'll only check more serious articles / deep discussions) to get personalized results.

It looks like it's progressing well.

One quick thing I noticed is that the thumbnails show up, even if there is no article for them: http://i.imgur.com/QrLZuxF.png

Hey, great job man. Not often a site distracts me for as long as this one did :)

In the initial tour, are the "top moments" hard-coded or actually being generated? Seeing "Aaron Schwartz Commits Suicide" in huge letters at the top of the column had me closing the tab rapidly.