Why is data a key ingredient to unlock the superpowers of cloud? Traditional computing is about taking a piece of information, an input, through an algorithm in order to generate an output. This can be a very tedious process. It requires someone to teach a computer what to do, one step at a time. Sure, it can follow the steps faster than any human, but it can only do what humans have programmed it for. Traditional computing therefore is not self-learning. With cloud computing, things change. You can provide input and output side-by-side and generate an algorithm. For instance, with image recognition such as Google photos, you can provide your photos, the input and identify a friend or family member, the output. Then you can generate a model that will learn from your photos and recognize that person even if their appearance goes through significant changes. That's the case, for instance, of a baby becoming a child and eventually in adult, whom the model will identify at every step of the way. In other words, when you have large volumes of data and lightening fast processing power, you can apply new forms of coding for the computer to become self-learning. This is known as machine learning. What does self-learning actually entail? At its most basic, machine learning emulates the human brain. Let me demonstrate. Suppose I show you a two column table with information about a house. In one column is a list of square footage: 1000, 2,000, 3,000 square feet, and in the second are the corresponding costs: $100,000, $200,000, $300,000. Your brain automatically starts to look for a pattern or a correlation between the two columns. Within seconds, it learns a formula. Square footage times 100 equals costs. In other words, your brain used input and output to produce an algorithm. Now if I ask you the costs of 2,500 square foot home. You're probably going to say $250,000 because you have that calculation ready. You can predict that cost, and recommend to save money for it. That's how our brain works. Decision-making is rarely this simple though, especially when it comes to buying a house. There are more and sometimes even complex variables involved. For example, what would happen if I show you a four column table that lists crime indexes and school ratings too, and then I ask you, "I have a 3,000 square foot house. It has a low crime index of 20 percent, and a stellar school rating of 10 out of 10. How much do you think it costs?" You probably couldn't give me the solution right away. Because even with four factors, four sets of data, you have to perform a complex regression analysis and it becomes way more difficult. It takes a good mathematician, and good software to actually do this, and we need to move from the printing platform to the IT platform to address this need. This is just with four factors. There are inherent limitations to the human brain's ability to take in non-linear data, analyze it, and make decisions based on it with a relevant level of accuracy. To make things more complicated, most people base their house buying decisions on many more factors. We use anywhere between 50 to 70 criteria most of which are actually unconscious. For example, we might have "a crush" on a house because the light of the rising sun was gorgeous in the living room, or avoid it because the cars in the street reflect a different social status than the one we are comfortable with. The point is, as humans, we most often make decisions based on our emotions, our experiences, and our memories without being fully conscious of it. This is where Cloud technology is truly differentiating. For example, a good machine learning model, learning from thousands of past examples and capturing data from photos inside and outside a house could predict that it has an 85 percent chance to generate a crush, and that this crush will be worth $20,000 with the 10 percent margin of error. This is why we say that the cloud is built to understand and map very complex and evolutive patterns such as human behavior. Machine learning emulates the processing and decision-making capability of the human brain, but without its limitations. This ability to combine tremendous compute power with vast amounts of data is what makes the cloud extraordinary, and why it will reshape every human industry. Unlocking the transformative superpowers of cloud technology to achieve your business objectives relies on leveraging the right data in the right way. As our real estate prediction machine learning model can take in and analyze more data, it's insights become more accurate. And quality also matters, which is why we relied on images to anticipate the likelihood of buyers having a crush on a house. This relationship between data strategy and the value of predictions is why data today is a form of currency. Most large organizations store a large amount of data on-premise, in the same way people used to store money in their mattresses. This makes data both vulnerable to attacks and unproductive. Instead, it needs to go to the "bank", not only because storing it there is cheaper and safer, but more importantly because this is where it will provide a return on investment, and the "bank" of data is the cloud. With the right combination of hardware, software, and machine learning capabilities, cloud technology enables you to ingest data from a variety of different sources in real time, analyze large volumes and different types of data, and create insights in seconds. As the volume and accuracy of data increases, algorithms become more precise, and insights are more meaningful. We'll look at how you can gain insights from your data when leveraging cloud technology while keeping data secure, private, and in compliance with third-party regulators in a later module. For now, let's look at a few examples where this extraordinary compute power and vast amount of data combine to unlock the superpowers of the cloud.
