Chapter 1

AI to Save the World?

Whenever I wish to deeply understand a topic I will always start with the bigger picture, as I believe it is critical to waterfall the “why” understanding along with the value chain. The why flows all the way from global trends, through to our workplace values, and finally to our daily work and life.


This book does not require any previous knowledge in machine learning or technical background, high school level math is more than enough. When a topic is worth delving into a deeper understanding for machine learning practitioners- then it will be labeled as “ML Pointer”. These sections are not critical and readers can skip them.  

The content is friendly to management, business, product, engineering, labeling, and anyone who WOULD like to learn more about the data principles behind AI. 

While the principles here apply to verticals like medical, surveillance, agriculture, retail, and many others, we will use mostly cats (and a dog) as our subjects. We don’t need more than that to explain the fundamentals and challenges behind the AI development process and show why bringing AI into production takes years. While we focus on image analysis as our example application, the fundamentals we will cover are valid for videos, audio, text, and other types of unstructured data. 

Why AI From 30K Feet?

Both the public and private sectors are going to invest heavily in AI, why?  

The world economy has major financial debt issues, beginning on the personal level (mortgages, student loans) up until corporate and government (bonds and social benefits). 

There are 3 possible ways you can resettle this debt: 

Since growth is the only happy way out, you can assume that it will be pushed heavily by governments, corporations, and individuals. Growth is simply the (inflation adjusted) increase of goods and services value generated by a given economy.  

We can describe the economic output as the result of following inputs: 

Let us look at these growth drivers. 


During the past decades we’ve exhausted our planet’s natural resources, with oil being a lead commodity. This path has come to a halt, we are close to maximum exploitation of natural resources around us, making the current situation sustainable both from inventory and climate is a challenge I hope we will pass. However, we can safely assume mother nature is no longer able to help us boost our future economic growth. The last century of oil-driven growth is over.  


With oil losing its prime status, a wave of clean cheap energy is rising. Cheap, renewable energy is good news for most and is going to help drive growth by reducing the cost of running machines (tools). It is hard to tell whether this transition is fast and smooth enough, yet as for now the cost of renewable energy keeps dropping which means the cost of automation is expected to drop along with it.  

Human population 

The human population is the main driver of growth, creating both the supply of products and services as well as the demand. On the global level, the more important aspect is the supply, we can only consume what we generate after all (noting that the way I see things supply is what powers the supply-demand growth cycle). Human workers affect growth in two aspects:  

  1. The numbers of workers/consumers 
  2. The economic value generated by a single worker.  

Technology’s main contribution to economic growth is by generating more things and services for less human work, driving costs of goods and services down.  

The value we generate as workers is a result of the actions we take during our work, there are two types of actions every one of us is taking while we work: 

While we can automate many of the human worker physical actions for decades (machinery and robotics), we are now entering a new era where many human cognitive actions can also be automated as well, bringing many businesses new labor efficiency opportunities, which tends to be a significant cost factor. 

If we take driving as an example: 

  • The required robotics exist for years (automatic wheel, breaks, gear, and gas control) 
  • Cars are becoming electric, renewable friendly.  
  • There is a significant global effort in creating an automatic driving cognitive application, a robotic driver (robo-taxi). The last part is the full automation of the human driver. 

Tesla is both smart and electric because its founder, Elon Musk understood these trends years before the market. Way earlier than other traditional car manufactures. Understanding the fundamentals and how technology is impacting them is key for creating great market timing. 

By now, the motivation behind AI is clear and powerful, it is fair to assume that without AI it is hard to forecast a positive outlook to our economy and standard of living.  

New Technology Brings New Challenges

It is worth mentioning that a significant social change is to follow AI progress, with Covid strongly accelerating it. These important and meaningful aspects are outside of the scope of this book, without discounting their impact or significance.   

AI is expected to bring significant social challenges: 

  • Inequality – from the workers who lost their jobs to machines and all the way to value distribution inequality, as of today the excess value(money) generated by this process goes only to a few tech giants. The process in which AI (e.g., self-driving cars) is taking main street jobs (e.g., the taxi driver) will repeat itself on every vertical in the coming decade. 
  • Fake news – The ability to generate fake texts, images, and videos will bring a trust issue, we will have a hard time recognizing truth from lies, while this is already happening today, what we experience is just the beginning of the process.  Loss of common trust will make us easy to manipulate as a society. We can expect social and governmental stress to increase in the coming years.  
  • Accumulated bias – Our own human biases will go into our smart agents through the data it is exposed to and will be amplified, many times unnoticed for a long time. This is a major risk in amplified weakening of already weakened populations 
  • Social dis-order – Governments are weakening, the power shift with data towards tech giants who gain the ability to manipulate billions of people, across nations in real-time. 
  • Taxation – Most of our tax system is built to tax the human worker, technology taxation is much more complicated. This will lead to public budget disputes, as economic growth does not find an easy path for public good growth (tax revenues). 
  • The gig economy – As out of the box jobs are disappearing (automated or crowd distributed), worker creativity becomes critical and many times leads to gig-workers, these workers’ social benefits status is not clear, and you can expect that the entire concept of career development and education institutions to go through major transformations. 

With the hope that governments, corporations, and citizens will resolve the above issues, we will focus on understanding the AI development principles moving forward.

Next Chapters

Chapter 2

I define a cognitive application as an application that can completely replace the collection of human cognitive actions, or the “thinking” part of a given work task or skill. In many cases these applications will start as human assistants, gradually replacing humans completely as they get more reliable and broader.

Chapter 3

So, data is critical for developing AI bots or cognitive applications, but that line of thinking can be misleading. The common phrase of “data is the new oil” is often used to express the value of data while ignoring the more important aspects of information and knowledge.

Chapter 4

AI development is essentially the process of collecting and organizing information. Data is collected, its meaning is extracted as information pieces, and then it’s structured into a format that allows future learning for the knowledge that these information pieces represent.

Chapter 5

The training process is the process in which we take our training set, i.e. the collection of data examples we’ve collected and create a model that learned from this example. We call this “training” and not “coding” since the model is created automatically from our data, with no coding involved. The result of our training session is a code we can then run that predicts its learned properties as a result of the new data.

Chapter 6

While often bias and variance terms are usually being discussed by data scientists and ML experts, understanding them requires no technical skills and is critical for anyone working with data-driven products, after all these are the data modeling bugs that will hurt our user’s experience and our product competitiveness. Time to gain deeper intuition on these concepts, no worries, you will understand them without a single equation involved.

Chapter 8

It is very popular to talk about machine learning these days while ignoring the teachers in this learning process, time to discuss the machine learning process from its less common perspective - as a teaching process.

Chapter 9

We are preparing to launch our AI app. We have basic models that are functional, we agreed with the pilot customer for a calibration period that allows our models to adjust to the fresh data and the data interfaces (APIs) with customers have been defined. In this chapter we will dive into the preparation and planning needed for launching and scaling our app deployment.