Often the best way to learn about a new field — especially one that is both complex and noisy, like AI — is to watch people who are truly proficient in it. What do they talk about? What do they spend their time working on, what things do they ignore? What kind of music do they listen to?
HINT: Someone working in marketing is not going to learn much about AI from another person working in marketing, for the simple reason that most marketers don’t understand the fundamentals of AI: they know the marketing-speak version of AI, which looks a lot like bullshit bingo. The term “AI” is just another entry in the lexicon of ad agency word porn, a sexy-shiny thing to whip out during new biz pitches and client dinners, but ultimately signifying nothing.
To whom, then, should an autodidact turn in order to better understand where AI is really headed, and why industry pundits insist on saying ridiculous things like “artificial intelligence is eating the world”?
Below are five people I find particularly insightful on various AI topics, ranging from neural networks to self-driving cars, social robotics, natural language processing, and AI startup investing. I highly recommend paying close attention to each of these people in 2017.
(Note: at 5, this list is necessarily — hilariously — incomplete; my goal is to be concise and useful, not comprehensive. Plus, I like prime numbers).
Carla Echevarria, Design Lead at Google
A masterful presenter, Carla is unrivaled in her ability to clearly explain neural networks to a non-technical audience. After listening to Carla, you will understand how hierarchical thinking is different than linear thinking. You will understand why traditional programming is useful for teaching a computer how to recognize a specificcat, but utterly useless at enabling the computer to teach itself “catness.” You will understand the preceding sentence without having to re-read it six times.
Carla has a deep background in design, with previous lead design roles at MakerBot, Facebook, and R/GA. Part of her focus at Google is on non-visual design challenges presented by voice-only interfaces, such as Google Home.
Shivon Zillis, Partner at Bloomberg Beta
What LUMAscapes are to adtech (remember adtech?), Shivon’s landscapes are to the field of machine intelligence. As a Partner at Bloomberg Beta, Shivon is laser-focused on machine and enterprise intelligence, and her investments include Shield AI, Datalogue, and Domino Data Lab, among others. If you follow AI long enough, you realize that some of the most important work comes out of Canadian universities. Shivon is on the advisory board of the Machine Learning Group at the University of Alberta, a leader in reinforcement learning.
Yann LeCun, Director of AI Research at Facebook; Professor at NYU
When people talk about “personalization at scale” they are talking about Facebook. With more than 1.2 billion users, Facebook Messenger has replaced SMS in the Western World and become the testing ground for bots and conversational commerce. Sadly, most chatbots suck. This is usually due to a lack of contextual awareness or emotional tone-deafness. If anyone can change that it’s Yann LeCun, one of the pioneers of deep learning. In fact, Facebook’s public foray into AI began in 2013 when it hired LeCun.
Rumored developments at Facebook include speech recognition and enhanced emotional intelligence for its “M” assistant, the latter underscored by the company’s acquisition of Ozlo just three days ago.
Fei-Fei Li, Director of Stanford Vision Lab
(Full disclosure: Fei-Fei Li and I attended Princeton University at the same time during the late 1990s. The similarities end there).
One month ago, Quartz featured Fei-Fei Li in an article captioned “Stanford professor Fei-Fei Li changed everything.” Everything! Li is a rock star in the world of computer vision; in 2009 she pioneered a new paradigm in large-scale visual recognition called ImageNet; the annual competition that followed is credited as a major catalyst to the current AI boom.
In this podcast with Frank Chen and Sonal Chokshi of Andreessen Horowitz (where Li is the Distinguished Visiting Professor of Computer Science), Li discusses the current historical moment in AI, the advent of deep learning chips (GPUs → TPUs), and convolutional neural networks, among other fascinating topics.
Andrew Ng, Co-founder of Coursera, former head of Baidu AI / Google Brain
As the founding lead of the Google Brain team, former director of the Stanford Artificial Intelligence Laboratory, and lead of Baidu’s AI team of 1,200 people (that’s not a a typo), Andrew Ng knows a couple things about machine learning. His early work included developing one of the most capable autonomous helicopters in the world (yawn), and he’s also a pioneer of online education, getting Stanford courses online in 2008 and founding Coursera in 2012.
Head over to Coursera if you want to learn more about any of the topics discussed here.
I hope this short(-ish) list was helpful. Please let me know.