Similarly, for many years, NVIDIA was known for the GPU, which was prominently utilized as the main component in entertainment consoles. When NVIDIA first created the GPU, the company was originally focused on 3D game rendering. That same basic product was subsequently repurposed as a main component in the NVIDIA Drive PX system, which generates data that is used for cameras, radars, and other sensors that comprise an AI platform for some autonomous vehicles. The present utilization of the GPU in the building of a platform enabling self-driving cars is sufficiently removed from the original intent so as to be considered a historical non-sequitur. Yet, the combination of ideas constitutes the proverbial “tip of the iceberg” when considering future possibilities for this technology.
There is no program that could be developed to consider all the possible permutations of situations that an autonomous vehicle might face. Just factoring in weather and visibility alone would require infinite amalgamations and responses. Traditional rules-based AI, which is basically a set of linear arithmetic extrapolations, cannot deal with the complexity of decisions needed to be considered in real time by the autonomous vehicle. Today, these AI systems learn in much the same way a child does. It’s this ability to learn that is key to making a self-driving car possible. Machine learning leads to deep learning, which effectively constitutes the simulation of decision making.
This ability overcomes the obstacle of having to explicitly program every possible situation. Today, investments are being made to enhance systems and take that decision making to an even higher level. To quote an article by Tristan Greene on The Next Web (https://thenextweb.com/artificial-intelligence/2018/04/10/future-ai-may-hallucinate-and-get-depressed-just-like-the-rest-of-us), “Scientists believe the introduction of a hormone-like system, such as the one found in the human brain, could give AI the ability to reason and make decisions like people do.”
Cloud: A Gateway to Infinity and Beyond
History has taught humanity that technology investments become gateways into new avenues of use, at multiples of what we could have anticipated. Think of the many ways we now leverage the cloud in the technology we use day in and day out, from our smart watches to Amazon’s Alexa. For many years, the barrier to widespread use of robotics has been the cost and complexity of programming. Imagine the programming that would be required to teach a robot to deal with the complexity of the ever-changing conditions of a construction site. Teaching an automated system to play chess at the level of a grandmaster is a much easier task then teaching a robot how to clear a kitchen after a meal. However, it is reasonable to envision leveraging the technology that enables a self-driving car to make a worksite robotic helper.
The ability to process copious amounts of unstructured data very quickly, by combining the information into the proper context that enables the system to make “educated decisions,” allows for an unlimited amount of situational applications. Imagine a new class of marketing applications that take full advantage of a data stream and adjust real-time decision making based on changing conditions at the moment, or consider a self-driving car that feeds its data to the cloud, optimizing the time it takes for you to get to your destination based on the current conditions or knowing exactly the speed to take so that you arrive at every intersection with a green light in front of you.
Blockchain and the Cloud
The investment in the technology known as “blockchain” has resulted in the foundation for modern “cryptocurrencies.” Think of blockchain as a secure automated ledger. When all parties reach a transaction consensus, the ledger automatically records the event and that ledger becomes a link in the DNA of the transaction object, just as RFID tags are used today to track everything from pharmaceuticals to livestock. RFID technology and blockchain capabilities can result in a very powerful combination, and a gateway of applications yet to be invented, all of which can be supported by cloud technology.