Until last year, I hadn’t written a line of code since developing COBOL mainframe systems in the 1990s. As the CEO of a cybersecurity company that makes extensive use of machine learning (ML) in its products, I am interested in truly understanding how this transformative technology works. While I’ll never code for a living, the learning process has taught me more than just coding.
Open Source Opens Doors
I started by learning to code in Python, the most commonly used programming language in the ML field. From there, I made use of a number of open source resources including Jupyter Notebook, TensorFlow, and Keras. All of these powerful and comprehensive tools are freely available. Their easy access has opened the door to advanced machine learning technology for many – reducing the barrier to entry and as a result, accelerating the pace of innovation across many industries and communities.
Quantum computing is another field that is being enabled, accelerated and transformed by open source. The 2017 release of the open-source framework Qiskit by IBM has paved the way. Over the last few months, I used Qiskit to program IBM’s quantum computers. IBM currently makes Qiskit freely available for use via the cloud through the IBM Quantum Experience. Ultimately this raises interest and the bar for its entire quantum ecosystem.
When the CEO learns to code, it doesn’t just benefit the CEO. When working with our team, I’ve found that I am able to ask better questions – the right questions – of our engineering and product teams. The ability to understand technology at a detailed, hands-on level provides invaluable insight. Even though I’m no coding expert, learning to code has given me the context to better articulate the benefits and competitive differentiation of my company’s products. Looking toward the future, I feel I will be able to better assess upcoming technology and the implications for strategy and investments.
As an added bonus, I believe that there is power in leading by doing. If a CEO can do this in his or her spare time, anyone can.
Impact on Cybersecurity
The impact of machine learning on cybersecurity is readily apparent. Traditional cybersecurity approaches cannot handle the volume and complexity of new threats. With machine learning we can build dynamic models that continuously learn from the characteristics and behaviors of malicious activity across millions of devices worldwide. In this way, we can keep up with cybercriminals who are also building increasingly sophisticated attacks using machine learning.
While it’s always enjoyable to learn something new and to better understand your technical team members, there is one last, crucial thing learning to code has given me: a fresh outlook as a leader. I feel my mindset has shifted to be more experimental and explorative, my creativity has expanded, and best of all, I’m more open to new ideas. It’s a great place to be as a leader.