We’re working on something <singsong>cooooooool</singsong>

Here’s what it looks like in our basement:

Secret Hello, Chair Project from Luke Iannini on Vimeo.

(Hint: what has 15,000 icons, can be comfortably cupped in your hand and makes a sound like “ba-ba-bum-ba-ba-bum-ba-ba-bum-bum-bum”? No, it’s not Dick Cheney.)

There are many impressive facets of the launch of the App Store, but there’s also no doubt it hit a lot of people in the head on its lumbering trot down the track. It has since started to raise its head and take a look around, and bit by bit, Apple has been repairing the damage, from allowing promotional codes to making some slight tweaks to their iTunes-based App Store to separate paid from free apps. But these have been incremental improvements rather than the radical changes needed to effectively deal with tens of thousands of apps.

Perhaps the biggest persistent issue has been app exposure - a week ago, the controversy was approaching fever pitch over the crowding of the much-watched App Store “Top 25″ lists. Developers felt the lists’ unintended emphasis of $0.99 apps was severely crippling the economic viability of developing iPhone apps. And it continues; Layton Duncan is compiling a litany of the many criticisms and suggestions being made by the developer community and its followers.

For example, Gruber, Casasanta and Ryu suggest price-weighting the Top 25 list, which would indeed help the economics at the top of the charts, but wouldn’t do much for the 9,975 apps out there sitting all lonely, stuck in a tiny hotel room, hoping for a better life. Universal Top 25 lists (weighted or not) are fun, but just like the Billboard charts, won’t help you discover Bob Dylan until he makes Blowing in the Wind.

The issue with such lists is they abstract the entire activity of a very complex market into a single vector: total sales. But there are now over 13 million iPhones and who knows how many iPod Touches in the world - that’s millions of people with millions of personalities and interests and tastes. To represent the subtleties of all those people as a single number does not show adequate respect for human complexity. You can prove this to yourself by heading to the Top 25 list right now and counting how many apps you think you’d buy. I got about 8, meaning I’m only 30% similar to the virtual anti-übermensch personality created by the aggregation of all customer desires. And I doubt if anyone actually wants 100% of what The List wants, because it is, again, a pastiche of many diverse groups of people clumsily summed together. Why not create a market-like system that observes these intricacies? That’s called a recommendation engine in today’s parlance, and we haven’t seen one yet. So we built our own, called Appalanche.

And to get this out of the way, here are our selfish motivations in creating Appalanche. One: each app you buy through Appalanche gets us 5% of the purchase price of that app, thanks to Apple’s affiliate program. Two: we learn lots of interesting stuff from people using our recommendation engines, that we can apply to improve Appalanche and any other product we develop using those engines (including Adpinion). And three: we like the press.

So, those are the selfish motivations, and they all rely on lots and lots of people using Appalanche. The only way to do that is if Appalanche is actually useful, so you see, it’s extremely important to us to make Appalanche an incredibly great app recommender. Which, fortuitously, happens to be a selfish motivation too…

That is: we wanted a way to input “I like Twitterrific, Tweetie, PCalc, and Deep Green” (all exceedingly high-quality apps), and for the response to be a recommendation of, say, Classics. Because, from what I can gather, people like me, that like high-quality apps, tend to like other high-quality apps. Sometimes, I’m not even interested in the app’s function! In the case of Deep Green, I bought it solely to play with its phenomenal UI and learn lessons about what makes a beautiful app so pleasing. I barely know how to play chess.

And that’s exactly how a recommendation engine works. It makes inferences. It won’t show you something merely because it exists in the same category, or sells at the same price point, as an app you love. It learns that, okay, the bunch of people that all like this app always seem to like this app too. What’s the relationship? We don’t actually care. We just know that when someone new comes along, and they like the first app, there’s a very high chance they’ll like the second one too.

And that manifests itself in all sorts of great ways. If someone’s not afraid to spend money on apps, they’ll naturally get recommendations of more expensive apps, because all the apps bought by people with enthusiastic wallets cluster together. If someone loves TripLog/1040, they’ll see other apps with a similar design philosophy, because all the apps bought by people who like to use apps with tons of immediately-adjustable parameters cluster together, too.

And, so the theory goes, if your app is not in their cluster, you didn’t really want to show your app to them anyways, because they’re not your customers!

Even better, recommendation engines are self-correcting. If it turns out that the aforementioned TripLog/1040-loving customer is in fact your customer, too, then there must be others like her. And so a new cluster will form, for people that like your subtle and elegantly designed apps, but who also need a little TripLog/1040 every once in a while (perhaps, when they’re stuck in a tiny hotel room, hoping for a better life).

So what it comes down to is this: James Thompson asked for “a store sorted by quality, not by popularity. Is that so hard?“. But, it is kind of hard, because whose definition of quality do you use (don’t answer that)? So, instead, we’re suggesting an App Store sorted by qualities - finding the apps sharing the most qualities similar to your own personal favorite qualities.

…

Um.

Using a recommendation engine. And, you know, not just some run-of-the-mill recommendation engine; you’d want a Really Great One that’s been heavily developed for years. And you’d probably also want a - okay, try out Appalanche already.

P.S. Incidentally, the biggest request we got last week when we launched Appalanche 1.0 was for a way to just tell it some apps you already know you like in order to “seed” the recommendations. And boy were they right! As of yesterday, we’ve added search to Appalanche, and we now recommend searching and voting on 3 or 4 of your favorite apps before you do anything else!

P.P.S. We actually anticipated this request, and there has been a Secret Experimental Feature in Appalanche from day 1 (we forgot to tell anyone about it : ) ). Simply take screenshots of your iPhone homescreens (using the “Home+Lock buttons simultaneously” trick), and email them to apps@appalanche.com. We run proprietary image recognition to figure out what apps you have and send you a link to a pre-populated Appalanche. This is still “experiental” because it is not 100% reliable yet - search is.

So, as you may have seen, we’ve recently launched our secret project, dubbed “Appalanche“. What it iz, iz, an iPhone App recommendation engine, optimized for iPhone Safari (press iPhone Safari’s “+” button to add it to your iPhone’s homescreen, where it will emulate a native app).

If you don’t, like we do, have a large cardboard thermometer that you paint red as the App Store ticks upwards in apps, or, if you have one, but didn’t notice that the Thermometer Painter recently cleverly painted the top of the thermometer bursting in a mercury volcano because the App Store recently passed 10 Fricken’ Thousand Apps, well, I’m here to tell you: the App Store now has over 10 Fricken’ Thousand Apps.

Here’s the problem with having 10,000 apps: finding 1. That’s a problem for developers and users alike; Apple has done very little to help users find apps. The solution they’ve presented thus far is twofold: one’s the new “iPhone Your Life” page, where Apple’s staff curate a selection of nice, broadly appealing apps, and give those apps the full Apple Marketing treatment. The “Featured Apps” pages in the iTunes and iPhone store serve a similar role.

The second solution is the famed and ignominious “Top 25″ list: one for “Most Downloaded (Recently)”, and one for “New Releases”. As any developer will tell you, there’s a massive difference between being “on the list” and not: you’re either flying high or drifting in obscurity. The “Downloads” list caused developers to race towards the bottom in pricing, ’cause a ton of downloads at $2.99 is better than no one even finding your app at $14.99. And, for niche (but awesome!) apps, this was never a solution at all. A niche app just can’t compete with “Bejeweled 2″.

The “New Releases” list was actually a positive influence at first: updating one’s app bumped it to the top of the list, encouraging developers to update their apps. But, naturally, cunning dicks exploited this by releasing a stream of insignificant updates to their apps. Apple quickly used the blunt fix of making the “New Releases” list only for entirely new apps, perhaps encouraging devs to just release a stream of crap apps instead.

As for the users, they’re pretty powerless. Even in the categories, there tend to be around 1000 apps each. How are they ever going to find the finest calculator, or RjDj, Mandelbrot, or Android FX?

Anyway, it all leads me to a salient quote by Netflix CEO Reed Hastings:

“I think that once you get beyond 1,000 choices, a recommendation system becomes critical”.

We’d like to propose a new model for finding interesting apps: tell us the ones you know you like, and we’ll find others that people like you liked too. It’s called: Appalanche.

The ego is perhaps the most impressive structure in the universe, simply because: it is capable of convincing its owner of their importance in the face of the most compellingly obvious evidence to the contrary, namely, the infinity accessible by peering upwards at any given moment.

If you need any more evidence, consider who or what is responsible for this very blog post’s thesis…

Great news! We may be hearing back from our intragalactic friends as soon as 2012. Hisashi Hirabayashi was nice enough to send along a note to the Altair system, only 16 light years away, in 1983 (the same year Magnum, P.I. started getting good, so hopefully they’ll be able to pick that up too).

And who wouldn’t respond to this?

(via Pink Tentacle)

“The dream of creating a color music for the eye, comparable with auditory music for the ear, dates to antiquity…”

— William Moritz via the iota center

I think that’s an interesting thing to ponder. I suppose my theory on why recorded visual stimulation has not taken off in the same way as recorded aural stimulation is that the former cannot yet be presented ambiently. I expect that once we have dynamic wallpaper or persistent volumetric holography gas, this will change.

Anyway, here’s a low-res 2-dimensional experience that I enjoyed, in bed (eeeew).

Dreamscape, by Vibeke Sorensen and Rand Steiger

Two recent “I can’t believe it’s JavaScript/Canvas” examples:

Cubescape

and Processing.js by the mystical John Resig.

I just want to tie together two ends of an Internet Story. A year and a half ago this TED talk made the rounds:

It’s a supa-neat video; Hans Rosling created some very compelling visualization software with apparent passion. It was noted at the time that his software had been acquired by Google, which made me say “hm, what’s that about?”. Was it for Evil Prognostication? Um, perhaps, but now you can too: they’ve released it as part of their Google Visualization API.

(via Ajaxian and All The Hip Nerd Blogs from a year and a half ago).

Here’s a video on one of my favorite topics: our good friend the Singularity. Michio Kaku (the affable string theorist) presents, which means that it’s on the lighter side but makes up for it in breadth. One of a three part series. And this comes again from Smashing Telly, an entire blog focused on exceptional documentaries. I can’t recommend it highly enough.