AI is like a sci-fi movie: it’s not new, but it’s really loud now

The term AI (Artificial Intelligence) is now as overused as the term blockchain was a few years ago. It’s newness is a mix of reality and hype, but there are some interesting signals that can be discerned through the noise.<br><br>I worked for the robotics division of one of the world’s biggest defense contractors in 2000, and they were working on practical AI solutions at that time. It wasn’t called machine learning, but they were using data to improve decision-making. They were already working on autonomous demolition trucks and autonomous drones. <br><br>I worked on some AI projects over the years, but I stopped doing any real work in AI in 2005. Here’s what I learned since:<br><br>* It’s really practical now. It’s not just concept, prototype, proof-of-concept, or anything like that. It’s a viable, deployable technology.<br>* It works very well for certain things, but very poorly for other things. For example, I was able to create a model that could predict the outcome of a complex scientific experiment (like making medicine) with a high degree of accuracy. That same model was utterly useless for predicting the outcome of a multi-player game of poker.<br>* The potential is huge. It can replace many jobs. It’s probably not a coincidence that the loudest voices against AI are coming from the tech companies that will be impacted the most. That might also explain why some of them are investing so heavily into it.<br>* Many people are afraid of AI, but it’s not sentient. It’s not human. It’s just a tool. It might be used for nefarious means, but that’s true of many other tools, like hammers, cars, and computers.<br>* Many people are fascinated by AI, but it’s not new. It’s been around for decades. We’re just finally seeing some fruits of the labor of scientists who worked on it over the years. It’s a sci-fi movie that’s finally made it to the big screen.