Technological Drivers of Modern AI

Besides advances in new conceptual approaches, theories, and models, AI had some other important drivers. Here’s a look at the main ones:

• Explosive Growth in Datasets: The internet has been a major factor for AI because it has allowed for the creation of massive datasets. In the next chapter, we’ll take a look at how data has transformed this technology.
• Infrastructure: Perhaps the most consequential company for AI during the past 15 years or so has been Google. To keep up with the indexing of the Web—which was growing at a staggering rate—the company had to come up with creative approaches to build scalable systems. The result has been innovation in commodity server clusters, virtualization, and open source software. Google was also one of the early adopters of deep learning, with the launch of the “Google Brain” project in 2011. Oh, and a few years later the company hired Hinton.
• GPUs (Graphics Processing Units): This chip technology, which was pioneered by NVIDIA, was originally for high-speed graphics in games. But the architecture of GPUs would eventually be ideal for AI as well. Note that most deep learning research is done with these chips. The reason is that—with parallel processing—the speed is multiples higher than traditional CPUs. This means that computing a model may take a day or two vs. weeks or even months.

All these factors reinforced themselves—adding fuel to the growth of AI. What’s more, these factors are likely to remain vibrant for many years to come.