Apple and Artificial Intelligence: The Odd Couple

Apple Builds Premium Computers as the World Moves toward Commoditized Invisible Computing

Apple is doomed. I said it. Sell your stock now, get out while you can.

Well, that isn’t quite true, but I do think that as we move into the next computing paradigm, the Cognitive Era, Apple’s business model and culture will need to change. The Cognitive Era describes the next phase in technology development in which the world around us The future success of Apple will not be determined by the iPhone, iPad, Watch, or even a car. For Apple to continue its dominance of the technology market and expand into healthcare, transportation, and fashion, it needs to move away from being a computer company. This may sound like a ridiculous statement considering Apple’s position at the moment. But as the Internet of Things gathers pace with every device connecting to the Internet, the value is in making these devices do smart things. Similar to electricity today, computing power will be on tap whenever and wherever we need it. The next computing paradigm, the Cognitive Era, with ubiquitous and invisible computing power poses an existential threat to Apple.Apple is a company that makes money from people who are happy to pay extra for a computer (phone, tablet, or watch). The long-term danger to Apple’s business is that AI solves problems such as natural language understanding, computer vision, and behaviour prediction.

Apple is a company that makes money from people who are happy to pay extra for a computer (phone, tablet, or watch). The long-term danger to Apple’s business is that AI solves problems such as natural language understanding, computer vision, and behaviour prediction.

Artificial intelligence becomes the only product differentiator

Right now, Siri, Google Now, and Cortana are similar enough that the user might be unable to tell the difference in quality. Continued progress in combining deep learning with other machine learning techniques such as genetic algorithms and bayesian inference, together with data network effects, means the quality of applications will begin to diverge. Smartphones, wearables, and cars will be bought based on what they can do rather than how they look. Sure, design will always matter and there will always be a high end of the market, it’s just this will matter less for the products Apple sells. Apple needs to become an AI company.

Apple’s only true competitor in the smartphone space is Google, and at its core, Google is an AI company. Google’s guiding strategy is to do whatever is necessary to collect data to feed into its AI engine. Google Fiber, Loon, and Chrome are all products designed to get more people using the Internet and to leave behind more data. Nest, Waze, and Dropcam are services that generate vast amounts of data. Moreover, Google has the best AI experts on the payroll. Geoffrey Hinton and his team are pioneers of deep learning and have invented many of the most widely used tools. DeepMind, a 2014 acquisition, was recently on the cover of Nature magazine and is leading the progress in general AI. When it comes to data and talent, Google is miles ahead of Apple.

Apple is Waking up to the Threat of AI

In the last 6 months, Apple has attempted to address the AI threat by acquiring 3 leading machine learning companies: VocalIQ, Emotient, and Perceptio. VocalIQ is a UK-based machine learning company that builds voice user interfaces. Perceptio is a US-based deep learning company focusing on developing smartphone-based computer vision solutions. Perceptio’s solution is unique in that unlike most computer vision products, it allows smartphones to identify images independently without requiring access to cloud-based data libraries. The most recent acquisition was Emotient which is attempting to automate facial recognition and analysis. Their vision is to build emotionally aware technologies.

The technology behind VocalIQ will be useful for the Siri and the Apple TV teams. Emotient and Perceptio will help bring local intelligence to FaceTime, photos, and Apple TV. The local angle is important firstly because Apple’s core business is in devices not cloud like Google. The more processing that can be done on the device the better for Apple. Regardless of how the acquisitions will fit existing products, the fact that 3 AI companies with deep learning expertise have been acquired in such a short space of time shows Apple’s commitment to bolster its AI capabilities.

Apple’s Secrecy and Closed Approach Will Limit Its Ability to Succeed in AI

What made Apple so successful in the mobile era was its fully integrated approach to building computers. By vertically integrating the supply chain and building the hardware and software, Apple has offered an unparalleled user experience that enabled the company to maintain premium prices and high margins in an exceptionally competitive market. This vertical integration required tight control across the value chain to ensure the products and user experience are market leading. This focus on providing the best user experience created a relatively closed and secretive company culture.

This culture means Apple has not embraced the open-source community in the same way as Google, Facebook, and other leading AI companies. For example, the 2015 Neural Information Processing Systems (NIPS) conference was the largest annual gathering of deep learning experts. In addition to Google and Facebook, Baidu and Microsoft were present; however, Apple was not. Over the last year, as detailed in The Only Thing That Matters in Machine Learning is Data, Google, Microsoft, IBM, and Facebook have released their machine learning frameworks for free to the open-source community because in the machine learning field, only data and developers matter. This strategy collects more data and obtains more developers and PhDs using their tools to make it easier to recruit them in the future.

Apple is still to publish a research paper on AI, despite using the technology and that Siri is a leading natural language processor. The secrecy runs counter to the prevailing approach within the AI and, specifically, the deep learning community. The small talent pool is attracted to the open approach of Google and Facebook and even the new venture OpenAI that plans to release all research to the community. In fact, the speed of progress in the field is due in large part to the collaborative approach of researchers who have worked together in small laboratories and have taken their academic approach to the companies who hired them. Apple’s secrecy and closed approach directly impacts its ability to attract the best deep learning talent.

Can Apple Adapt to the AI Era?

The integrated approach that made Apple successful in the mobile era could be the very thing that makes it unsuccessful in the cognitive era. Apple is one of the few companies to have successfully navigated a paradigm shift in computing, from PC to mobile. That shift, however, favoured Apple’s focus on user experience and, therefore, a closed and controlled culture. In the PC era, the main buyers of PCs were heads of IT departments who decided what to buy based on quantifiable indicators such as processor speeds, RAM, and compatible software. Apple’s focus on the user interface is less tangible; therefore, its proposition was weaker. In the mobile era, however, consumers are the buyers of devices that are used pretty much all day every day. User experience is far more important. Apple has the right company culture to build consumer devices.

The shift from the mobile era to the AI era does not suit a closed company culture. In the field of AI, an open and collaborative culture is needed. Without changing the culture, Apple’s products will be weaker than its competitors.

AI is the biggest threat to Apple’s position as the most valuable company in the world.

Why Apple bought Beats

The Next Episode: Why Apple Wants to Buy Beats

With 2014 being the year of crazy tech acquisitions, Apple obviously doesn’t want to miss out. When Google buys Titan Aerospace or Nest, and Facebook acquires WhatsApp or Oculus VR, they are making bets on the future. Whether it’s drones, the internet of things, mobile messaging or virtual reality, each of these domains are clearly part of the future in some way and the value of buying early is not to be disrupted later. It makes at least some sense. The music business is anything but future technology. Music is no longer the key battleground, Apple won out with iTunes and has dominated the industry ever since. Music is no longer a lock-in service as music streaming services are platform-agnostic, meaning Beats would provide almost zero ecosystem value for Apple. This feels like more of a bet on the past, and at US$3.2 billion (the largest ever acquisition for Apple), a big bet. So what is this all about?

“They say rap’s changed, they want to know how I feel about it”

The pricing model for music is changing. iTunes unbundled the album and its a la carte model of selling music won out. There are signs that the subscription model is now overtaking a la carte. Music subscription revenue increased 50 per cent to US$1.1 billion in 2013, whilst downloads declined by two per cent in the same period, for the first time since the iTunes store launched. But Beats Music only had 111,000 subscribers as of March 2014, the paid-for figure is likely to be considerably lower if you strip out the 90-day free promotional trial offered by AT&T. Spotify has over six million paying users and Deezer has over four million. Beats Music, despite being newer, has a very long way to go. The key differentiator for Beat Music is its curated experience. Quality content discovery is still too complicated for most users and curation solves the “what to listen to” problem. Apple’s defines itself by its simplicity and user focus. Beats would certainly be more interesting to Apple than a Spotify or Deezer and a lot cheaper to boot.

“Got my red Dre Beats on, tryna put my peeps on”

With 64 per cent market share of the premium headphone market in North America, Beats was on track to record revenues of US$1.4 billion in 2013. Beyond market share, the most valuable part of Beats is the brand. Much like Apple products, serious professionals bemoan the poor quality specs and dismiss Beats as poor quality. That misses the point. Headphones are now for more than just listening to music; they are a piece of clothing, an item that can reflect personality. Brand, not sound, is key. Beats headphones are now ubiquitous, worn by football players, pop stars and actors; they are cool. Even cooler than Apple’s white headphones in the iPod era. That cool factor translates its revenues, as customers flock to pay a huge premium. Cool, premium and high margins; the business sounds a lot like Apple.

The Message

The proposed purchase is a music subscription play first and foremost. Apple needs to move to a subscription model and must feel that the iTunes organisation is not capable of changing the business model. The curated product and ease of discovery fits with the Apple ethos, and it has plenty of runway for growth especially if bundled with an iPhone contract in the future. Second, they get a profitable US$1.4 billion business, a strong brand, and a team that has intimate knowledge of designing products customers want to wear. All of which seems like an awful lot for a US$3.2 billion price tag.