-
Book Overview & Buying
-
Table Of Contents
Creators of Intelligence
By :
AA: We all too often hear of data science/AI projects having high failure rates. Do you think this applies to the APS, and if so, why are the failure rates high?
MM: Do we see a high failure rate? Going back to what I said before about experimentation in data science, start with “science.” The word “science” is in there for a reason. We have to be able to experiment. I guess it depends on why they’re failing. Maybe something failed because the problem wasn’t well designed; it may have been really important but just not solvable right now, and that does not necessarily mean that people are failures. It means you have to say, “OK. Put it on the shelf for a bit, let’s see what else we can do in the meantime, and we’ll come back to it.”
But if something failed because it wasn’t the right problem, then there’s a prioritization issue. I will not proceed with a problem...
Change the font size
Change margin width
Change background colour