Article: The Great AI Awakening in (New York Times Magazine)
The story of Google Brain’s path to a massive breakthrough in translation as a case study for the growth in efficacy and flexibility of neural networks.
As the world’s largest technology companies continue to hire and acquihire talent in this space, my sense is this story will become less the exception than the template. Research or processes designed with one purpose in mind that quickly find significant utility when applied elsewhere is a bet we’re seeing placed on a near weekly basis.
What I was particularly struck by in this article is the rate of progress possible when the right research or concepts are matched to a seemingly unrelated application. The key ideas to enable progress in this space were available long before they were applied, yet once the right parties at Google elected to experiment with them, significant performance gains followed in short order.
This is promising for our capacity to improve, but is frankly a sobering reminder of how little even our experts with resources know about the potential of our technologies before they are applied. Now it’s a pleasant surprise to be lauded, but as we experiment with higher impact applications of our ideas, we should be equally careful to consider unpleasant surprises.
