-- Must Read Links --
Artificial Intelligence Is Stuck. Here's How To Move It Forward. NY Times.
Gary Marcus, whose A.I. startup was acquired by Uber, makes some great points about the limits of current A.I. "microdiscoveries" and the gaps we still need to bridge to get to a solution. He encourages us to think bigger. I personally would love to see the U.S. government launch a Manhattan project for A.I., or something the equivalent of CERN. The massive investment would have huge payoffs.
The Data That Transformed A.I. Research, and Possibly The World. Quartz.
A look at the history of ImageNet, how it came to be and the immense impact it has had on the A.I., particularly the field of Deep Learning.
Why Teaching Will Be The Sexiest Job in the Future A.I. Economy. Intuition Machine.
A look at how future intelligent systems may not be programmed, or trained, but rather, grown. In a world where A.I. learns piece by piece, teaching these A.I.s could become incredibly important.
The Dual Use Dilemma In China's New A.I. Plan. Lawfare Blog.
A great look at how China is setting up their A.I. strategy such that much of their investment and innovation can serve multiple purposes, particularly military use cases.
How We Built a Virtual Scheduling Assistant at Microsoft. Harvard Business Review.
An inside look at a human-in-the-loop model to build a calendaring assistant. This is, in my experience, the top approach by most A.I. companies right now, which is why figuring out how to get annotated data is the top priority for A.I. companies.
The Future of Computing: When A.I. Met Moore's Law. Medium.
A fantastic post by Azeem Azhar about the cycle of more compute leading to more powerful algorithms leading to higher demand for data, which fuels the need for more compute.