MFML 014 - Human creativity in AI

you do see these creative applications like hey drunk island write me a poem and for those you don't really know what it means to write a poem well

you kind of hope for something good to come out of there but you're not sure what you're expecting well for those things you shouldn't really hand over

the task entirely to this system instead what you should expect is that it is just a way to inspire your human creativity it's just going to mush up some information for you

and propose something that looks like a poem that's just some way of combining a whole lot of data and you're going to look at this and you're going to actually use your own

judgment as the human creative and you're going to ask does this inspire me can we work with this and if it doesn't you're going to kick that island tell it hey good do this a different way over and over again

until something looks like it's worth pursuing and then you're gonna do that trick that all these other folks do which is you're gonna put that out there on the internet and pretend like that was the only

version ever look at the poem machine learning wrote for me machine learning writes poems says the newspaper heading it's actually you that's writing the

poem machine learning is just jumbling stuff up for you in some half-hazard way like monkeys on typewriters just a little better because it's finding patterns rather than doing it purely at random

but it's a way to filter human creativity that's how you should think of those applications but we're not quite sure what we're looking for there we have it heavily

interfacing with the human and the human is at both ends of the process