25 July 2016
Career choices are some of the most important ones we make in our lives. It is often pointed out that we spend most of our waking hours working, and, of course, how we manage our careers directly correlates to our earnings and standard of living. These results are relatively easy to calculate, as they are quantitative factors. I think it is just as important to consider the qualitative impact of career decisions. Our work is where we often find a sense of accomplishment and purpose. It shapes our perspective and how we view the world, which directly affects our creativity and ability to make an impact. So, as you can imagine, I believe each of us should make smart and deliberate decisions about career paths.
We often first learn about hot, new fields through the media. Predictions are made about high-growth areas and the skills needed to keep up with the expansion. Clearly, data science and machine learning are among the hottest fields right now. You cannot escape declarations that Data Scientist is the #1 job to pursue. Even back in 2012, Harvard Business Review declared Data Scientist the sexiest job of the 21st century!

The hype isn't limited to the prospects for careers in this field. The implications of machines being able to learn on their own without being programmed by humans are staggering. This opens paths to progress in most every aspect of our lives -- education, transportation, and healthcare, just to name a few. Every major tech company has invested heavily in artificial intelligence and machine learning. You've likely heard of some of the more talked about applications, like Google's AlphaGo (which beat one of the world's top GO players) or extensive development being done to bring us autonomous cars. 

There's certainly a lot of talk, but is it all just hype? It is true that many of these promised applications are not yet available to us in the market, but they soon will be. What some people don't realize is that machine learning is already here. Every time you ask Siri, Google Now, or Cortana to call someone or find information for you, your phone is employing machine learning to make that happen. Whenever Netflix or Spotify suggest TV shows or songs for you, they're using machine learning to identify what you'll like based on what you've been choosing to watch or hear. So, these technologies are already helping you move through your day more efficiently and effectively.   

Based on all of this excitement, you will hear many people telling you to jump into a data science or machine learning career. It seems people are always eager to give unsolicited advice when it comes to life choices. They are quick to suggest what you should do. It is usually based on their own experience, their own missed opportunities, or what they've heard everyone else buzzing about. Some of you may remember the sage advice given to The Graduate -- "Plastics!" It is wise to remember that many people have gone broke chasing "the next big thing".
Every major tech company has invested heavily in artificial intelligence and machine learning.
Share this

So, should you invest the time, energy and money required to build a career in a new field just because everyone says it's a sure thing? The answer is no. You shouldn't do it solely because of the hype. However, conventional wisdom is a good data point to consider in your decision-making process, and that process should be a rigorous one. I'd like to share with you my own assessment method that I've used throughout my professional life. It is one that has reaped great benefits for me. When assessing whether to enter a new field that has been touted as having a positive future, I look at factors that can enable or disable the realization of that promise. I create a side-by-side comparison of these factors to determine the relative power of momentum vs that of limitations. 

I first used this method in the 1980s when mobile technology was nascent. The first mobile phone went on sale in 1983, the DynaTac, for $3,995. (You might remember Gordon Gekko using it another famous film, Wall Street.) That price point was a disabler, one which would generate limited demand. That factor could be weakened, however, if the price of the components came down. At the time, I was working at Intel, where I became intimately familiar with the predictions of Gordon Moore, the company's co-founder. "Moore's Law" said that the number of components in an integrated circuit would continue to grow dramatically. I bet that this would be proven true, and that the prices would come down precipitously. It was a good bet. Mobile devices are now integral to our work and personal lives. In fact, it is hard for us to escape them now.

I conducted a similar analysis in the 2000s around video when smartphones appeared. It was clear that this larger phone screen would be able to accommodate video at least visually, but the disabler was limited bandwidth. In order for infrastructure companies to make the investment needed to increase bandwidth, there would have to be promise of a payback. In my assessment, the big enabler here was advertising. Given that more than $530 billion is spent on advertising every year and almost 40% of that goes to TV advertising (ZenithOptimedia, 2015), I bet that advertiser demand for a new TV-like screen would be big enough to warrant an infrastructure build out. As co-founder and CEO of Meru Networks, the technical solution to abating bandwidth constraints for mobile video also became clear to me, which only strengthened the case for enablers. This turned out to be another good bet. Advertisers are already spending more than $5.5 billion on mobile video ads annually and growth is expected to stay in the double digits through 2019 (eMarketer, 2016). 

You can apply this same methodology to the fields of data science and machine learning. Let's first take a look at the disablers. One of them is societal fears. Some people have voiced concerns about creating machines that can think for themselves, that can act independent from human direction, and that do not inherently have a moral code like humans do. Even prominent figures in the tech industry, like Elon Musk, have called for a cautious approach. But even those concerns are accompanied by an admission that there is no stopping it. Artificial intelligence is already here. 

A second disabler is the dearth of workers skilled in data science and machine learning. Companies are struggling to hire qualified people who can bring to life the new products and services that will incorporate the promise of artificial intelligence. However, you could see this as even more of a reason to enter this field. Demand for data scientists is high, and supply is low at this point.
With the spread of the Internet of Things (IoT), we are collecting unprecedented amounts of data about every second of our lives.
Share this

The enablers, on the other hand, are on fire! The biggest one is the proliferation of data. The more data we have, the more we're able to feed to machines to fine tune their learning. With the spread of the Internet of Things (IoT), we are collecting unprecedented amounts of data about every second of our lives. It might seem like this amount of data would be too voluminous to manage and make good use of, but that leads us to our second major enabler. 

The scale at which we can now store and process data in the cloud is one of the most important technological enablers of our lifetime. Because of this we have access to massive amounts of cheap storage. Data storage on AWS is actually free for up to 5GB and then costs only pennies per GB beyond that. This coupled with ever-increasing processing capabilities, a one-trillion fold increase over the last six decades, has propelled and will continue to propel us forward.  
A big enabler for not only the technology itself, but also for related career paths is the proliferation of courses in data science and machine learning. The number of courses being offered to people looking to develop skills in this field is growing every year. As you comb through the offerings at Coursera, Udemy, Udacity and others, you'll see many of their most popular courses are data science-related. Even two years ago, the largest class at Stanford University was a graduate-level machine learning class

When you take the time to run the analysis, it becomes clear that the scales are tipped in favor of enablers for careers in data science and machine learning. In my experience, the more successful paths in life are the ones where there is an abundance of factors that can remove obstacles. That's not to say that any path is a certainty, but I believe you will make infinitely better bets when making informed decisions instead of using guesswork. It has certainly worked for me. 
Tap to read full article