Day 31–100 Days of ML — Automating Comedy
Ah, yes. Another lofty goal where I attempt to do the impossible just on sheer grit and willpower. If you’re into ML and stats, you’re probably familiar with the story of the stats student who rushes in late to class and copies the board thinking that’s the homework. S/he returns to class stating, “This was harder, but I did it,” only to discover when the bemused professor tells them, “These were unsolvable.”
I thought that’s how this comedy automation process would go down and I’d create trillions in opportunity and get to be a guest on Siraj’s YouTube channel and there would be much rejoicing.
Nay, that is not how this went down in the slightest. My first mistake was in the broad terminology, “Automating Comedy”. My next mistake was one I made several years ago when I decided Statistics was not a great place to pick up women. Yes, young Jimmy had whacked priorities that have led to a diminished tech tree, but I courage and I go on.
Looking back at my styles and modalities of comedy, I came up with six:
Stand-up
Improv
Sketch
Sitcom
Cinema
Songs
These are still too broad for the task at hand, but I think if I piece these down further and really hammer in a solution that has a distinct input and output, then I can throw AI, Machine Learning, and Deep Learning at the problem and scale out humor.
On the next post, I’m going to write about the Improv game, “What are you doing?” See you there!
