Following Apple’s Worldwide Developers Conference in San Francisco, The Register published a piece that claimed their team had a drink every time the tech giant mentioned machine learning at the event. The end of the headline tails off into drunken mistyping, which is as strong a review of Apple’s eagerly awaited conference as you will find. The words were mentioned together so many times that you could hardly fail to identify just where Apple’s priorities lie.
An announcement that has gone slightly under the radar at WWDC is the reveal of Core ML, a programming framework that Apple hopes will make the iPhone a more fertile ground for machine learning models. The platform will come as part of iOS 11, and will give developers the tools to ‘load trained machine learning models onto an iPhone or iPad and then use them for generating insights inside applications,’ as described by VentureBeat. Developers have been able to do this for some time, but Core ML is designed so that apps can process data locally using machine learning, negating the need to send user information to the cloud. Google announced a similar framework at its I/O developer conference last month, and it seems machine learning is the next major battleground in the iOS/Android war for supremacy.
This wasn’t the only part of MMDC to focus heavily on machine learning, though. HomePod is also as much about Siri as it is music. The limelight given to the clunkily-titled piece of hardware seemed a little excessive for what Apple are keen to describe as a speaker rather than a virtual assistant. Perhaps the most impressive tech is, indeed, the speaker’s spatial awareness. The company demonstrated how the device can detect the room in which it sits and its relation to it, from which it can adapt its sound to best fill the room. The tech is impressive, but with a price tag of $349 it ought to be bringing something new to the table.
The company claims that the HomePod is Apple’s attempt to straddle both quality audio and smart technology, but it’s the former that it will be hoping catches on. It is, at its core, a way for Apple to get people to engage with Siri in an intuitive, novel way, more because it’s there rather than because its capabilities are so necessary. It takes Siri out of the pocket and into the home proper, an omnipresent assistant to rival that of Amazon and Google.
Some have viewed the HomePod as Apple simply catching up with its rivals in terms of voice-assisted hardware, while others see it as a trojan horse attempt at sneaking Siri into as many homes as possible. The problem with this is that Apple is positioning Siri as a ‘musicologist’, a smart speaker rather than a truly smart assistant. At $349, users will be expecting more than just music. Apple tends to have no problem selling its hardware, but with other popular (and only slightly less pretty) alternatives on the market at half of the price, it’ll be interesting to see just how well HomePod does.
Between Core ML and HomePod, Apple has well and truly thrown its hat into the ring on machine learning. Siri is a perfectly fine virtual assistant, but the tech giant will be hoping that HomePod will makes its capabilities more readily available in the home and ergo more widely used. WWDC was dominated by Apple’s designs in the machine learning space - you don’t need a drinking game to see that.