Ajay Agrawal’s welcome message from the 2016 conference:

“Machine intelligence impacts such a wide range of economic activities, including transportation, agriculture, healthcare, energy, retail, and manufacturing, among others, that it is difficult to fathom a single economic feature common across all implementations. However, there is: a drop in the cost of prediction. The recent advances in machine intelligence represent a staggering reduction in the cost of prediction. This is meaningful because prediction is an input to a host of activities, and when the cost of an input falls so precipitously, there are two foundational economic implications.

1. Expanded use to new applications

First, not only will goods and services that use prediction as an input become cheaper, but we will begin to use prediction as an input for things for which we never previously did. As a historical example, consider semiconductors, an area of technological advance that caused a significant drop in the cost of a different foundational input: arithmetic. Activities for which arithmetic was a key input, such as data analysis and accounting, became much cheaper. However, in addition, we started using the newly “cheap arithmetic” to provide solutions to problems that were not traditionally arithmetic problems. An example is photography. We shifted from a film-oriented, chemistry- based approach to a digital-oriented, arithmetic-based approach. Other new applications for cheap arithmetic include communications, music, and drug discovery.

The same goes for prediction. As the cost of prediction falls, not only will activities that were traditionally prediction-oriented, like inventory management and demand forecasting, become cheaper, but we’ll begin to tackle other problems for which prediction was not traditionally an input. Consider navigation. Until recently, autonomous driving was limited to highly controlled environments such as warehouses and factories where programmers could anticipate the range of scenarios a vehicle may encounter and program if-then-else type decision algorithms accordingly (e.g., “If an object approaches the vehicle, then slow down”). As a result, it was inconceivable to put an autonomous vehicle on a city street because the number of possible scenarios in such an uncontrolled environment would require programming an almost in nite number of if-then-else statements.


Inconceivable, that is, until recently. Once prediction became cheap, innovators reframed driving as a prediction problem. Rather than programming endless if-then-else statements, they instead simply asked the AI to predict: “What would a human driver do?” They outfitted vehicles with a variety of sensors – cameras, lidar, radar, etc. – and then collected millions of miles of human driving data. By linking the incoming environmental data from sensors on the outside of the car to the driving decisions made by the human inside the car (steering, braking, accelerating), the AI learned to predict how humans would react to each second of incoming data about their environment. us, prediction is now a major component of the solution to a problem that was previously not considered a prediction problem.



2. Changing values of other inputs

When the cost of a foundational input plummets, it often affects the value of other inputs. The value goes up for complements and down for substitutes. In the case of photography, the value of the hardware and software components associated with digital cameras went up as the cost of arithmetic dropped because demand increased – we wanted more of them. These components were complements to arithmetic; they were used together. In contrast, the value of lm-related chemicals fell – we wanted less of them.

All human activities can be described by five high-level components: data, prediction, judgment, action, and outcomes. For example, a visit to the doctor in response to pain leads to: 1) x-rays, blood tests, monitoring (data), 2) diagnosis of the problem such as “if we administer treatment A, then we predict outcome X, but if we administer treatment B, then we predict outcome Y” (prediction), 3) weighing options: “given your age, lifestyle, and family status, I think you might be best with treatment A; let’s discuss how you feel about the risks and side effects” (judgment); 4) administering treatment A (action), and 5) full recovery with minor side effects (outcome).

As AIs improve, the value of human prediction skills will decrease because machine prediction will provide a cheaper and better substitute for human prediction, just as machines did for arithmetic. However, this does not spell doom for human jobs, as many experts suggest. That is because the value of human judgment skills will increase (complements). We’ll want more human judgment. For example, when prediction is cheap, then diagnosis will be more frequent and convenient, and thus we’ll detect many more early-stage, treatable conditions. This will drive up demand for treatment and emotional support, which is provided by humans. Overall, the value of prediction-related human skills will fall, but the value of judgment-related skills will rise.

The conference

So today, as you hear the presentations, learn about new applications, and experience the magic in the demos, we encourage you to reflect not only on the state-of-the-art of this technology but also on the trajectory of our shared future as these implementations develop, machine-based predictions continue their unrelenting march forward, and our judgment-related skills become ever more important for the development and prosperity of humankind.

We will also briefly transition from immediate applications of narrow AI into a longer-term discussion of general intelligence, which transcends prediction and contemplates understanding. A few expert practitioners, not just science fiction writers and futurists, are convinced that eventually “AIs will compete and cooperate with us just like other people, but with greater diversity and asymmetries. So, we need to set up mechanisms (social, legal, political, cultural) to ensure this works out well. Inevitably, conventional humans will be less important. Step 1 – Lose your sense of entitlement. Step 2 – Include AIs in your circle of empathy.” (Professor Richard Sutton, University of Alberta).

Here’s to a productive conference.”


About the Creative Destruction Lab
The Creative Destruction Lab is a 9-month long program for massively scalable early stage technology companies. The program revolves around meetings with high profile entrepreneurs, and investors –  G7 and ML7 Fellows and Associates.  Ventures accepted to the CDL get the chance to meet one on one with mentors every eight weeks for objective oriented sessions that allow them to gain valuable insight tailored to the success of their companies. The ultimate goal of the program is to maximize equity value for CDL Ventures.


Rotman School of Management,
University of Toronto

105 St George St,
Toronto, ON M5S 3E6

Directions »