Re:Making AI algorithms doesn't require a PhD (Score 1)

+1 this stuff weakens the whole industry -- I just got through losing a huge debate in which a family member expressed frustration over what she perceived to be irresponsibility in the field of AI research. I felt that researchers were doing just that: creating proof of concept that demonstrates refinement of technique, while she felt that they had some responsibility for making sure the output was racially unbiased, safe, complete, accurate, etc. I agree that these are important goals, but I feel that there's a huge divide between model R&D and production usage of a trained model. Turns out that it's really difficult to make a convincing argument to someone whose only exposure to ML is through articles like this, that blur the lines between research and production by claiming to put AI in the hands of everyone. AI is only ever going to be in the hands of people that take the time to understand it, regardless of how many pretty interfaces are put between the data and the user. Everything else is going to be buggy and frustrating to work with.