Many have predicted that AI has finally reached a breakthrough period after cycles of AI winters. My intention to do this research back in the summer was initially inspired by one A16Z episode. The episode basically covers the history of AI and how it possibly accumulates to current stage of AI, potential breakthrough. In my opinion, the breakthrough coexist with the potential bursting of tech bubble around this area due to the monopoly by tech giants.
To form mature opinion about a subject I barely understand, I first needed to get my head around what exactly is Artificial Intelligence. As shown below, the top row is concerned with thought processes and reasoning, whereas the ones on the bottom address behavior. Also, the definitions on the left measure success in terms of human performance, whereas the ones on the right measure against an ideal concept of intelligence, which we will call rationality.
So it’s safe to say that AI research is the research of intelligent agent that can act and think humanly, and, act and think rationally. To constitute such agent, one needs to design a program to map percepts to actions, and architecture, the computational devices for programs to run on.
Through years of research, the focus has also shifted over time. In the book, Intelligent Technologies for Information Analysis, it is discussed that more focus has been on building on existing theories than proposing brand new ones, to base claims on hard experimental evidence rather than on intuition, and to show relevance to real-world applications rather than toy examples. Such proposed idea made sense to me. It definitely explains the chart below and current AI ecosystem which I will later expound on.
Once having an idea what I got myself into, I started looking at the industry stats:
- Investment in AI technology has risen from $1.7bn in 2010 to $14.9bn last year. So far in 2016, as of 06/15/2016, over 200 AI-focused companies have raised nearly $1.5bn in equity funding, reaching an all-time high deal activities in this field.
- More and more VCs and industry giants like Google, Twitter, Salesforce dived into this field through either investment or acquisitions. CBinsight has done a great deal of data visualization on the M&A activities, you can easily check them out to see what I’m talking about.
Now the ecosystem, which I definitely to come up with all by myself. One article on Medium by Bloomberg Beta partner Shivon Zilis helped me a lot. One way to look at the AI Ecosystem is to look at them by application sectors: Agent, Autonomous System and Enterprise; the other way is to look at them by industry verticals such as Healthcare, agriculture and etc. There’s also a third set of players who sell their technology expertises as products.
The three sectors I want to focus on are Agent, Autonomous System and Enterprise sectors. As I mentioned before, AI research has pivoted its focus on providing solutions to real-life problems, so naturally smarter communication, smarter driving and smarter business are the top aspects people would love to improve on.
Some interesting insights/thesis that I have drawn through the research are:
- Autonomous Systems (Robotics) will remain as the major growth driver in AI industry
- “Time to Her” (TTH): With the boom in messaging platforms, we are expected to see more investment in Virtual Agents.
- The market will also see great growth opportunities in Enterprise AI.