Comment Worry about using, not creating AI. (Score 4, Insightful) 121
Just back from WWDC17 and I have this takeaway: leave the creating/training/designing/refining machine learning models to the academics and companies with deep pockets. You're not going to catch up with the PhDs that have a head start on you, especially without a unique authentic problem at hand that nobody else is working on yet.
Instead, USE the models that exist. Maybe train 'em with new/different data if you feel compelled, but mainly learn what models exist (natural language processing? Sentiment analysis? Image recognition? Speech recognition? Real-time identification of objects in video?). Learn how to use those models to solve the problems you're working with, or another team is dealing with, or that isn't even being considered for technology and humans are still doing it. The PhDs will keep creating new and better building blocks, just like we started out with basic web tools and now we have WebRTC. Our jobs will be to apply them. And that requires a lot less linear algebra. I think we can all say amen to that.