Gigaom AI Minute – July 1

Transcript

Bloomberg Businessweek recently reported that the phrase "artificial intelligence" is occurring more and more often in earnings calls. Just a few years ago it wasn't mentioned hardly at all, and now in the most recent quarter there have been hundreds of mentions of it.

Now, to be sure, this is warranted to some degree. Artificial intelligence makes breakthrough after breakthrough after breakthrough, and it's frankly difficult to keep up with them all. But the reason the phrase is so easy to work into a corporate narrative is because "artificial intelligence" is a term that does not have an agreed upon meaning. This stems from the simple fact that "intelligence" is a term that does not have an agreed upon meaning.

At its simplest, an AI is something that responds to its environment, but this is an incredibly low bar that applies to your sprinkler system as it's able to detect when the yard is dry, or the cat food dish that feeds the cat when the food gets too low. A more relevant definition is a system that responds to a dynamic and changing environment. In other words, a system that learns as it goes.

Hershey recently announced a system that I think qualifies. In the past when they made Twizzlers, the rope like candy, they either had to overproduce to make sure they got enough in the bag to match or exceed the weight or they had to weigh the bags repeatedly over and over again, but recently Microsoft Azure Cloud was employed to track 60 million data points in the manufacturing process so that the system can adjust in real time and make the exact right amount of Twizzlers. This really is a real-world legitimate application of AI in the broadest sense of the word.