The release of Apple's iPhone 4s gives us all pause - analysts and pundits have heralded it as just a bit faster with a nicer camera but nothing else. Indeed, the hardware has undergone just the typical evolutionary cycle that we have come to expect from these refreshes. But what most have not realized is that the software is a much bigger deal in this case.
The introduction of Siri, the artificial intelligence personal assistance, was smartly orchestrated and expectations were well managed by Scott Forestall and Phil Schiller. Phil suggested that it was beta software but the demonstration was very smooth. This is the first time in computing history that AI (Artificial Intelligence) is embedded into an operating system – this is a big deal! Today's kindergartners will grow up in a world where they dictate commands to their personal mobile computing devices - things such as "show me the location of the nearest candy store" making all these utopian Star Trek scenarios reality (remember Start Trek IV? Scotty speaking into the computer mouse as if it was a voice gateway to the computer). It is probably just a matter of time before Siri will find its way into the next release of OS X and the other iOS devices and once again, Microsoft and Google play catch-up.
Now something completely different yet related - I was so furious with AT&T that I switched my wife's iPhone to T-Mobile knowingly foregoing the ability for 3G data. But here's the thing: in Manhattan it's actually very rare to have 3G data (on AT&T). In fact, most times you can be happy to get any data. So changing to T-Mobile and running on the slower EDGE network does make sense in this situation and repeated heuristic tests (personally performed) have shown that there was not much of a difference between 3G and EDGE, with the exception of having cut my wife’s bill in half. (Disclaimer: I will switch to T-Mobile as well and no, she was not a guinea pig, it just happened that we needed a new contract for her). But this illustrates what the last mile problem represents, namely lack of capacity. For the cell carriers, it is a balance between marginal benefits and cost increases - or in other words, they aim to maximize the marginal increase in capacity for every cell tower. The issue is that he carriers don't really make their money with data - this is still a voice game for them and data is just an add-on that serves its purpose for the carrier - or at least, the current business models would hint at that given the various pricing tiers still measured in number of voice minutes.
Siri for all its wonders requires a data connection to the cloud in order to work best. Without knowing how Siri is built internally, I would think that some basic understanding of grammar is done on the phone in software - possibly tied to a particular voice (although Scott has denied that in the demo) but in this form, it will not go beyond simple commands such as scheduling meetings, setting alarms and do Internet lookups as demonstrated. Siri needs the cloud where it has a tremendous amount of data stored that it uses to figure out what has been said and what is relevant in a particular context. So questions like, “what is Apple’s market capitalization if the stock price increases $10” will have to be handled by the cloud. This is probably not different from IBM's Watson - the first system to beat two Jeopardy champions. Watson was able to do it because it had an arsenal of data available that it could access in real-time.
This is where the last wireless mile is becoming a real showstopper. The wireless carriers know that long-term everything is data and charging by the minute (or seconds) is a legacy model from the old Ma Bell that has been around for a good 100 years now. There has to be a paradigm shift when it comes to data as it will be rapidly more important for Smartphone users than voice. Not just because of applications like Siri but because voice itself is data and even today I can use Skype or various other VoIP solution on my iPhone.
Apple finds itself at the mercy of the wireless carriers that are trying to maximize their investments in infrastructure by pushing voice minutes onto their customer base. Apple could care less about telephony, after all the one component in the iPhone that is not revolutionary is the telephony component. In that sense, this is a story of limitations by the supplier. One possible remedy from this is to vertically integrate the supply side. However, that would require Apple to get approval to operate as a wireless carrier and it would require heavy investments. Given Apple’s balance sheet, the investment in infrastructure would be doable but that would also mean limitations in terms of distribution of handsets. Apple makes its money with the handsets – the more carriers that have the iPhone in their offerings, the more Apple will sell iPhones. If Apple were to become its own telco, it would potentially limit its reach because it would compete directly with other carriers and they would not be very happy and might drop Apple – especially given the rapid rise of Android as a real contender.
The last mile problem is not just an Apple story – Google has a similar problem with Android and its cloud offering. Both companies have made it very clear that the cloud will play an important role in their respective companies future. However, Google is a step ahead because that is Google’s primary business and so Google has already plans to solve this problem: enter white spaces.
In the US, the mandate to switch TV broadcasts from analog to digital, freed a wide range of radio frequencies that are now available as white space – or as unlicensed device spectrum. Google, along with Microsoft, Dell and HP and other names, is a member of “White Spaces Coalition” which aims to use the available white space frequencies for super charged WiFi (with the FCC’s blessing). Out of all the coalition members, only Google appears to have a complete value chain – that is, it has every piece of the puzzle to offer an end-to-end wireless Internet solution. Not only does it already have the cloud, and the client software (Android), it is also in a position to build wireless hardware devices: routers and terminals (aka. phones) due to the acquisition of Motorola. It’s funny how most people have commented on the Motorola deal as something whereby Google gets patents – I think it goes much further than that and it might usher in a new area of wireless computing that bypasses the traditional carriers and/or extends the traditional carriers with supercharged WiFi. If that were indeed the case, and we would see Google offering new devices that could be installed by municipalities as well as new handsets that could be offered directly through Google’s website, it would also be a wakeup call for the existing wireless carriers and that would in effect also benefit Apple. I’m almost tempted to say that Google might become Apple’s savior – if not, it might just be a question of time when Apple joins the “White Spaces Coalition”