I write columns on news related to bots, specially in the categories of Artificial Intelligence, bot startup, bot funding.I am also interested in recent developments in the fields of data science, machine learning and natural language processing ...
I write columns on news related to bots, specially in the categories of Artificial Intelligence, bot startup, bot funding.I am also interested in recent developments in the fields of data science, machine learning and natural language processing
There is growing concern in some quarters of the media about the misuse of the term artificial intelligence as a mere buzzword, as little more than the newest way to market a company's products or image, an issue I discussed in an interview in leading Spanish daily El Pais after participating in an OECD forum on the topic in Paris last June.
Make no mistake: artificial intelligence in general and machine learning in particular will be one of the major drivers in technology in the coming years: the ability of an unsupervised algorithm to learn from data will allow us to do things beyond our current understanding when it comes to tackling stable, highly predictable problems and where it is reasonably easy to collect large amounts of data.
At the same time, let's be clear: we're not talking here about machines capable of thinking as such, rather the ability to create products that make a competitive difference for companies with the right managers. Who are the right managers? Definitely, not the kind that believe that artificial intelligence can simply be bought and installed, or worse, those who see it as a marketing buzzword.
Those will soon come up against a harsh reality: any development of this type entails long, manual and complex processes of goal definition, data collection and transformation, along with feature engineering before any results can be expected. Using a neural network or tagging does not make you an artificial intelligence company, much less so overnight.
Even though many of the barriers to entry have gone down thanks to advancements such as Machine Learning As A Service (MLaaS), it is still a long process, and you should better start soon.
An interesting example is the development of competitive dynamics in the smartphone segment: obviously, one of the main factors that will define leadership lies in the nature and performance of the chips used. The announcement by one of the most interesting players, ranked third by market share, Huawei, that announced it will be using a new artificial intelligence chip (or, more properly, System-On-a-Chip, or SoC) just a few days before Apple launched full screen in its event its A11 Bionic microprocessor was no coincidence, nor was Google's acquisition of the assets and staff of HTC: a $1.1 billion operation that will allow it to incorporate artificial intelligence into smartphones, turn it into a differential and extract money from it.
So what is an AI chip? Basically, a marketing label. The Bionic in Apple's A11 is just a commercially attractive name to something as boring as a chip, but indicates something very important: the development of more powerful chips and some specialized functions to power perform certain tasks related to machine learning. Basically, being able to provide the right processing capability to, for example, process a dot matrix in real time to identify a face, or manage a better voice assistant. The Bionic bit may be marketing, but the benefits of the chip are potentially very real, and point to a smartphone ecosystem where manufacturers will compete to develop and use more competitive chips.
On the one hand, we have all the manufacturers that incorporate Qualcomm chips, which are now the almost de facto Android benchmark. At the same time, Google wants to dominate the voice assistant market and is preparing to incorporate a new generation of proprietary development chips specifically designed for these types of tasks, putting the company in a position to take on traditional chip makers and generating a debate over which artificial intelligence tasks will take place on the cloud and which ones on the device.
Apple, meanwhile, intends to go into battle with its A11 and with algorithmic developments focused on Face ID, Siri and photography, products where the user can clearly see how the application of artificial intelligence is progressing, while Huawei is set to unveil its new chips and advances in its next model, the Mate 10. Beyond that, there is not much more to say: it has never been easier to access competitive and powerful chips for practical artificial intelligence applications in handheld terminals. Increasingly, a terminal's attractiveness will be decided by the application of artificial intelligence to all kinds of tasks: we will take our photos with the help of an assistant based on this technology, we will talk with a voice assistant who will understand us better and better, another algorithm will recognize our face even if we grow a beard or use glasses, and several more will take care of things yet to be dreamed up. In short, thanks to the development of artificial intelligence, the marketplace is about to get a whole lot more competitive. And that is not hype or buzzwords: that's reality.
What businesses need to do now is think about where the smartphone industry is headed, how its competitive dynamics are being shaped by artificial intelligence, and then use that as a case to reflect upon the many lessons and consequences for their future, in their industries and in their products. And if you can't think of any, then be afraid, be very afraid