Software used was Torch-RNN (https://github.com/torch/rnn), and it's trained on a corpus of romance book titles from the King County and Seattle Public Libraries.
I made another bot. This one is a neural network trained on 30,000 romance book titles. Its "creativity" setting ranges between .5 and .95, since those were producing funny (to me) titles.
I wonder how many other bots are powered by @acb's dada engine? I'm guessing not a lot, but it's a pretty good tool for that.
I made a bot: https://botsin.space/@WarehouseWorker
It toots four times a day about the things you find as you go about your rounds in the Warehouse.
@rjl20 Does this work?