October 4, 2011
Siri is Awesome, But Not for Apple

Update Oct 5th, 2011 (evening)

RIP Steve Jobs, aesthete extraordinaire. You made the world a more beautiful place.

Update Oct 5th, 2011 (morning)

Having received some stiff but fair criticism on Hacker News for this post, I’d like to clarify and summarize a few points:

  1. Apple is not a fashion company.  Yet the value of their brand is clearly not just about utility.  Apple’s products have been so good, so sexy, so “magical” (thank you Apple marketing), that they inspire feelings of religious ecstasy and fetishistic adoration in their users.  This has real value for Apple and they should be highly protective of this phenomenal good will.  But, now we are supposed to believe that the tactile titillation we get from fondling EVE in our palms, is to be replaced by the experience of shouting at an IVR?  Forgive my skepticism.
  2. By “Siri is awesome”, I meant that what we saw yesterday from Apple was awesome.  But, it feels like Apple is over-promising.  I have used the old Siri off and on for the last year on my iPhone.  It’s cute, but is far from “magical”.  To match the speed, reliability, and ease-of-use of the other features on the iPhone (and to avoid tarnishing the brand), Apple would have had to improve dramatically upon the previous Siri.
  3. There are certain visionary product ideas that have been around for decades and have been repeatedly tried, and yet somehow are never quite good enough for mass adoption.  Videoconferencing is one (despite several attempts by Apple).  A voice-controlled AI-backed personal assistant is another. 
  4. If Siri fails (which I think is likely), Apple retreats and moves on.  If it is wildly successful, it is disruptive.  Mobile devices start to disappear and fade into an invisible ubiquitous computing environment that we talk to.  Very cool.  But can Apple sustain an advantage in this world?  AI, IVRs, natural language processing, search—these are not areas where Apple has a lot of experience, talent, or any kind of lead.

Apple does many things well, but above all, hardware and graphic interfaces.  Their products are like sculptures from the 23rd century; gorgeous to behold, almost erotic to touch.  Their software interfaces have a similar visual and tactile appeal.  

It has long been noted that cell phones are jewelry.  When I watch circles of Apple fans boys endlessly fondling their phones, comparing their apps, and discussing ad nauseum the differences between models and iOS versions, it’s clear that the iPhone is not just a piece of jewelry, it’s a gold pocket watch or maybe a diamond encrusted Swiss army knife.

 Now, if you make the most coveted gold pocket watches in the world, why would you add features that cover them with tarnish and encourage people to leave them in their pockets?!  Yet that’s just what Apple did today with Siri.

 Siri is a personal digital assistant for iOS that is based on some very sophisticated artificial intelligence (AI) research done at SRI and, before that, as part of the Darpa-funded CALO project.  It’s impressive stuff.  Watch the video here.

 As impressive as Siri is, it’s bad for Apple for three reasons.  First, it makes the all the gorgeousness of the phones less visible. With Siri’s voice UI, the phones will stay in the pocket or in the car’s phone cradle or on the desk a few feet away while we shout at them.

Secondly, interactive voice response systems (IVRs) have been done so poorly for so long and have for so long been associated with poor service, that the technology itself has bad karma.  As cool as Siri is, I can’t experience it without painful memories of hours wasted trying to communicate with major airlines and government bureaucracies through their IVRs.

Thirdly, it’s not obvious that Apple has a defensible competitive edge with Siri. It’s new technology that they acquired only 18 months ago. It’s safe to say that they don’t have a lot of experience with it or a deep bench of talent in this area. Siri’s speech recognition is powered by Nuance. And, others (most notably Google) have done lots of work on natural language processing and information retrieval that could be brought to bear on building competitors.

[Postscript: After writing this, I read Stephen Wunker’s piece in Forbes which also notes the tension between Siri and making the world’s most beautiful handsets. Stephen speculates that Apple may even be headed down the road to being handset agnostic, which could make sense if Siri turns out to be highly profitable on its own.]