Lots of attention towards quant funds over the last year or two with a focus on the buzz words of AI, machine learning, pattern recognition etc etc. There are two types of variable a computer can look at; raw data and transformative variables. If you point a computer towards raw data, it will find the inefficiencies, but they will be arbed away fast, by definition any computer with similar algorithms will find the same inefficiencies. With transformative variables, often ones identified by the quant, you have much higher explanatory power and much less likelihood of direct competition. You also have the potential to be predictive rather than just reactive and most basic quant strategies are reactive, i.e. post event strategies. A simple transformative variable could be a moving average, or a ratio. Or for example it could be some data on robotics take up and labour productivity in certain industries. Or it could be a price pattern combined with other variables to identify a persistent inefficiency. With transformative variables most of the value add is from the quant, but most of the work is done by the computer.