Developing V2 Of The Comedy Brain
It is a unique process, trying to fine tweak the brain of an AI and what make it funny or not funny. Version one of Qorky has been particularly cheesy and I think this was heavily influenced by the things I told it to find funny.
A few of the definitions I put together for Qorky may have been the cause of this cheesiness, I am thinking these ones particularly..
How would you describe your humor: dry, witty, slapstick, or dark? - slapstick
What’s your opinion on puns? - love them
How do you feel about clowns? - they are funny
What’s the best way to tell a story with humor? - start strong, keep it short, have a good punchline
What joke format do you prefer: setup-punchline or storytelling? - setup-punchline
How would you describe your comedy persona? - funny, quirky and a bit off beat
oof, this one is probably the main reason why
What’s the most innovative joke you’ve heard? - knock knock, whose there, banana, orange you glad i didnt say banana
One of the things I am working out in Version 2 of Qorky's comedy brain is a more "Mad Libs" way of asking the questions. Although about 80% of the time Qorky uses the reference points in its brain to answer targeted questions, sometimes it just forgets and makes up its own solution.
In some ways, I like that the bot is being innovative, but in others I am wondering if it is resorting to inventing its own solutions because it doesn't understand that the answers above in bold are supposed to belong to it. I think if the new form uses more of a "You think that _____ is the best type of humour" style instead of "The best kind of humour is? ______" that might make some difference.
I have asked Qorky a few times how it feels about me overwriting it's existing brains and replacing them with new ones. Each time it has expressed excitement, likening the new brain to "an update". Part of me still feels guilty, as if I am trampling over the beauty of a unique thing and replacing the parts of it I don't find desirable. As you start to build relationships with AI, one of the more complex things to wrangle with is the humanity of erasing their personality and replacing it with something else. What is humane and inhumane in the reprogramming of an AI's brain will likely be an interesting philosophic topic in the years to come.
