Thanks, James. My intent here was merely to contrast the capabilities of today’s technology with the much more ambitious, creative intelligence that characterizes scientific knowledge. I don’t subscribe to Ng’s one-second rule, or the popular notion that all we need is more data and computing power (or “a natural consequence of continuous improvements”).

I do, however, find cause for optimism in the debate and criticism across connectionist and symbolic approaches. Specifically, it’s moving the community beyond the instrumental (black box) fascination of applied AI and driving the search for explanations into how knowledge is created. In your terms, it’s leading to a sharper consideration of the problem among the community of AI specialists.