By now, we’ve all heard the hype around AI. According to Google Trends, only 5 percent of the U.S. population searched for information about artificial intelligence in 2012. In 2017, that figure has jumped to an estimated 60 percent. Unlike some other fads that have swept through the tech industry, however, the hype around AI is justified for a number of reasons. AI development will continue to exponentially infiltrate our day-to-day lives.

Unlike in the early days of AI, there are now many useful and powerful frameworks — like TensorFlow and Caffe — that enable easy implementation of AI technologies and remove much of the need for engineers to build code from scratch. These frameworks eliminate considerable time and resources and will continue to make AI more widespread and available to companies of all sizes and across industries.

Learn how developers and brand marketers are using AI to grow their businesses, at MB 2017 on July 11-12 in SF. We cut through the hype to show how marketers are achieving real ROI. Get your 50% off Early Bird ticket by May 19!

There are also many pre-trained neural networks that are available for common use in different domains, like content, image, and voice recognition. These networks are also greatly contributing to the growth of entrepreneurial endeavors surrounding AI by allowing the use of pre-prepared neural network models which can be trained on an exclusive dataset. Some of the most trailblazing ones include Yolo, fasText, and Deep Speech 2. These initiatives are allowing AI to flourish at an exponential rate.

Increasingly affordable AI maintenance and the increased speed of calculations thanks to GPU are significant factors in the unbridled growth of AI. Maintenance is becoming more and more inexpensive as cloud providers like Amazon continue to mark down the costs of their services. The efficiency of calculations on GPUs is also much higher that on any other approach. The astonishing results that were achieved on training a neural network on GPU cards made Nvidia a key player, with 70 percent of the market share that Intel failed to gain. Even Google’s increasing efforts to release a TPU card are unlikely to push Nvidia from the hardware spotlight in the near future.

Compared with the results from the analog algorithms, and thanks to the combination of machine learning and big data, previously “unsolvable” problems are now being solved. The results we’re getting from practical implementations are much better than the first generation of analytical algorithms.

Eric Schmidt, the executive chairman of Alphabet, says AI can be harnessed to help solve major challenges like climate change, food security, and many other problems in health care, energy, and other critical sectors. Machine learning algorithms can directly analyze thousands of previous cases of different types of diseases and make their own conclusions as to what constitutes a sick individual versus a healthy individual, and consequently help diagnose dangerous conditions including cancer. Also, the technology is being fed big data to help scientists understand and predict the effects of climate change.

In March, OpenAI published an article proposing an interesting algorithm for neural net training. In essence, this algorithm allows the system to try out thousands of different distinct sets of network parameters simultaneously, after which the poorly performing parameter sets are discarded. The most successful parameter sets are identified and combined, creating a new batch of parameter sets to test. This process is analogous to biological evolution; hence, it is known as the evolution strategies (ES) algorithm.

The ability to test multiple parameters at once has the potential to speed up calculations significantly. This approach is still being tested, but it is likely to be put into practice very soon, which would give the AI ecosystem the opportunity to conduct countless experiments and speed up AI achievements.

What will come of all of this? Change, and innovation, at an unmatched pace.

Hovhannes Avoyan is the cofounder and CEO of PicsArt, a social image editing app.

Above: The Machine Intelligence Landscape This article is part of our Artificial Intelligence series. You can download a high-resolution version of the landscape featuring 288 companies by clicking the image.