I like: Josh Tenebaum MIT lecture on building machines that see, learn and think like us

Reading Time: < 1 minute

One of the better presentations on the current state of machine learning that I have seen. There is the interesting academic / commercial divide in terms of likely outcomes. Industry isn’t going to create that general purpose artificial intelligence that everyone is writing about. The academic research is very far way as well. In many respects the over promise and under-delivery would be familiar to AI research pioneer Marvin Minsky. Minsky lived through a number of AI booms and AI ‘winters’ when the hype was deflated.

Despite Tenebaum’s reality check on the industry we still see commercial AI concept films created to sell clients on the future promise of a general purpose AI. Agencies and technology companies doing this are only hastening the rise of another AI winter.