The era of big data with a simultaneous increase in computing power in recent years has propelled us to faster and greater analysis through Artificial Intelligence (AI). The year 2019 will see this trend continue especially given the expansion of data collection through streams such as IoT.
There are four trends which shed a light on the current BBC forecast of “the year of data”, namely, data storytelling, AI development, cloud computing and hardware trends. We also saw the start in 2018 of a potential change in the use of mobile computing for AI.
Skilled data-based storytelling will grow as businesses leverage omnipresent data collection to satisfy their customers with useful personalised insights. A popular example is “Spotify 2018 Wrapped” which feeds back to customers a year’s personalised history and recommendations. This hunger for storytelling goes beyond usual analysis and is supported by the need and demand for machine learning (“ML”) and deep learning (“DL”) skills. This is not made easy by shortage of people with the right skills: for example, in the jobs market in London the number of jobs offers is twice the number of Data Scientists available.
Secondly, the current development in AI tools has led to an increased variety of software which offer automatic ML/DL infrastructures, enabling the use computing power to compensate for lower skills in AI deployment, by using special models which choose and optimise the most relevant solutions for the user. Yet this should enable more non-IT people to benefit from AI in making informed and faster decisions.
Computing-wise the “data year” has to be viewed from two perspectives, scalability through cloud (hardware on demand) and hardware market itself. The popularity of cloud solutions, where AI models are typically deployed, is growing due to increased demand, pushing seemingly more AI products and their faster development & deployment (scale factor). It goes in hand with the demand for Cyber Security related to the new data streams.
On the other hand, revived competition in the PC market challenging Intel’s dominance has been stimulated by the growth of AMD’s new line of products in 2018. AMD announced a new line of powerful processors built in 7nm technology directly competing with Intel’s’ ones projected at 1/3 of competitor price. Hence soon, we can expect a drop in costs for computing delivering AI applications.
Another major and unexpected shift could potentially affect the AI market with a silent innovation happening in the backyard of personal computing, in the vast “mobile computing market” of smartphones. Low-power low-cost devices enabling AI on the fly may reshape the market going against the centralised idea of cloud computing relying on the bandwidth speeds for full utilisation.
© CONTEXT 2020