Acquiring AI expertise is both a technical and emotional journey

Last week, Microsoft invested $1 billion in OpenAI to support its mission to safely usher artificial general intelligence (AGI) into the world and work together on AI technology for the Azure cloud platform.

Just over three years old, OpenAI is known for its quest for the AGI holy grail, creation of state-of-the-art AI like GPT-2, bots that beat top Dota players, and attracting investors like Elon Musk and Sam Altman. That’s why it was kind of a surprise this week to hear CTO Greg Brockman describe in detail the challenges he faced when growing from being a capable programmer into a machine learning practitioner.

He doesn’t offer answers sufficient for every programmer or business executive interested in picking up machine learning, but it’s an honest, personal account about encountering mental challenges while he was increasing his technical understanding in machine learning.

The most helpful part of the post might be that Brockman acknowledged how it feels to be a machine learning novice “overwhelmed by the seemingly endless stream of new machine learning concepts,” as well as feelings of frustration, a lack of confidence while others excel, and feeling “half blind.”

Brockman said he was aware that programmers with a background in linear algebra can be machine learning practitioners in a matter of months.

“Somehow I’d convinced myself that I was the exception and couldn’t learn. But I was wrong even embedded in the middle of OpenAI, I couldn’t make the transition because I was unwilling to become a beginner again,” he said. “You need to give yourself the space and time to fail. If you learn from enough failures, you’ll succeed and it’ll probably take much less time than you expect.”

He also mentioned that a new personal relationship, where he felt like he had the support necessary to feel safe to fail, helped him on his journey.

Brockman wrote in particular about this transition process within OpenAI. With billions of dollars in support and plenty of AI researchers on staff, it’s a very unique place to learn far different than an average business where there’s likely a dearth of AI expertise available to pull from. But he described a road that a great number of business leaders will have to travel as their organizations implement AI into their companies.

“There are many online courses to self-study the technical side, and what turned out to be my biggest blocker was a mental barrier getting OK with being a beginner again,” he said.

A lot of people have to cross this bridge. With the rise of machine learning, both technical and non-technical courses are growing in demand. Compelling recent examples include the AI for Everyone from Coursera, Udacity’s nontechnical AI program for program managers, and Microsoft’s free AI Business  School for executives. There’s also Amazon’s Machine Learning University, which received 100,000 sign-ups within its first 48 hours, AWS VP Swami Sivasubramanian said last month at Transform.

A KPMG report and survey released earlier this year found that more than half of business executives plan to implement some form of AI this year. Both the KPMG report and a Microsoft business executive survey found a correlation between company performance and ability to  projects. So yeah, learning how to apply AI can mean a need for technical prowess and capable understanding of subfields like reinforcement learning or deep learning, as well as an understanding of how it can reshape company culture.

Machine intelligence is being applied across society and is being called essential to business as well as modern-day citizenship. So everyone has to start somewhere, but obtaining the knowledge required to become a regular user of machine learning isn’t just about being intelligent. To succeed in becoming an ML practitioner can require patience when you feel half blind, fortitude when you don’t excel as quickly as your colleagues, and the need to hurdle unanticipated emotional barriers.