Yep, we’re in total agreement here. I think a lot of AI projects tend to get over-ambitious in their claims that they can create chatbots or analytics systems that can be “everything to everyone” but end up creating things that just aren’t all that useful.

I think there’s a lot we can learn from the success of the Bach AI, especially in the social space where the user experience’s is all that matters. Music that writes itself in any genre is a futile endeavor, but it’s possible to pass the Turing Test if you’re looking to imitate one composer in a specific genre, style, and time period. It might sound counter-intuitive, but the more limitations you put on an AI, the more likely you’re develop something resembling a “personality” of sorts, which actually makes the machine appear to be much more human. I think that even in the age of big data we shouldn’t be afraid to draw boundaries and limitations since that’s where we’re most likely to see interesting results!

Thanks for reading!