Yeah, really our main constraint was the 24 hours, and getting this to MVP stage. Really interested in getting more feedback though. If you take a look at the comments under http://lesswrong.com/lw/e26/who_wants_to_start_an_important_... with my username and you can read a little about how this kinda developed. We're still trying to figure out whether there's a compelling and sticky feature set that we can build a meaningful, mainstream service around.
I think there is potential here. As I mentioned above, Anki has a few quirks that make it cumbersome to use on occasion. My top three complaints, accompanied by potential solutions, would be as follows:
01. Getting information into Anki.
You've already got this covered pretty well. The ability to clip notes from around the web or add them via other methods is wonderful. Think of a tool like Evernote or Clip.to
02. Poor Quality Decks.
Let's face it, finding a good pre-made deck in Anki can be a challenge. You can easily solve this, simply offer your users a variety of high quality decks. The better the quality of the decks, the more value. It can take a REALLY long time to make a good deck, I'd gladly pay a modest fee for a well made deck that saves me hours of prep time.
03. Bad Design & UI.
I love Anki, but she isn't the prettiest girl in the room and she isn't always easy to get along with. A little bit of elegant design and thoughtful UI would go a looong way in boosting the perceived value of your product.
01. Totally agree with this. I hate the inertia to adding new info.
02. The best decks are always the ones you create, since they have personal value.
03. Ankidroid is the OSS android version and it's awesome. I bought the paid iOS app to support the author/service, but IMO it's really ugly and unusable (but then I think most iOS apps are unintuitive and crippled so take it with a grain of salt)
I've been interested in this for a long while especially in memory models for forgetting patterns beyond simplistic SR expanding schedules.
We developed http://membean.com that utilizes some of this technology and we've been able to make this a successful business albeit in a restricted domain. But as others have pointed out .. it's hard to sell this. We've managed it because of stellar content built around our engine: http://membean.com/exemplars. Without excellent content it would have been hard to get this off the ground despite us able to show that we outperform existing SRS tools.
You bring up a good point. Perhaps targeting a specific demographic would make the sell a bit easier?
I don't think it occurs to many people to seek out and use memory tools. However, there are groups of people who do have a very specific need for these tools and actively seek them out. I'm thinking of people like medical students, language enthusiasts, etc... Anyone who has an immediate need to learn and recall large amounts of information.
By intentionally limiting the scope of what you offer, you can better target prospects and tailor your offer to fit their needs like a glove.
As you pointed out, the quality of decks is important. Limiting scope allows for creation of quality decks. At the end of the day spacing is just one aspect of forgetting - strong encoding is the other and for that you need rich varied content. Creating quality content is hard work. We started off with a focus on algorithms and engines and then realized that even if we slap the very best Memory engine if you have crappy content - no one cares (and retention is marginal over simplistic flash cards). We invested a lot of work on content and varied encoding and it's paid off and we are now branching into different verticals.