Large Language Models (LLMs) have shown significant limitations in understanding creative content, as demonstrated by Hessel et al. (2023)'s influential work on the New Yorker Cartoon Caption Contest (NYCCC). Their study exposed a substantial gap between LLMs and humans in humor comprehension, establishing that understanding and evaluating creative content is key challenge in AI development. We revisit this challenge by decomposing humor understanding into three components and systematically improve each: enhancing visual understanding through improved annotation, utilizing LLM-generated humor reasoning and explanations, and implementing targeted alignment with human preference data. Our refined approach achieves 82.4% accuracy in caption ranking, singificantly improving upon the previous 67% benchmark and matching the performance of world-renowned human experts in this domain. Notably, while attempts to mimic subgroup preferences through various persona prompts showed minimal impact, model finetuning with crowd preferences proved remarkably effective. These findings reveal that LLM limitations in creative judgment can be effectively addressed through focused alignment to specific subgroups and individuals. Lastly, we propose the position that achieving artificial general intelligence necessitates systematic collection of human preference data across creative domains. We advocate that just as human creativity is deeply influenced by individual and cultural preferences, training LLMs with diverse human preference data may be essential for developing true creative understanding.

During his 20 years as the New Yorker's cartoon editor, Bob Mankoff developed an interest in the creative potential of artificial intelligence. In 2005, he helped found the magazine's cartoon caption contest and his desk began receiving between 5,000 and 10,000 entries a week. Mankoff – who studied experimental psychology at university – worked with Microsoft, and Google's DeepMind, on projects that attempted to develop algorithms to distinguish between funny and unfunny submissions. For tech firms, developing machines with a sense of humour makes commercial sense. As electronic assistants and robots play an ever greater role in our lives, we'll want them to be good company.

"Innovation," Jeff Bezos once said, "happens by gently lifting a grandfather and asking him for six different ideas." It's the work of Botnik, a new AI-assisted humor application that scours various types of human-created, word-crowded content--from season-three Seinfeld scripts to Yelp reviews to Bezos' shareholder letters--in order to build predictive, idiom-specific keyboards. Those keyboards, many of which are available on Botnik.org, The best Botnik creations, like this PBS-derived set of otter facts, retain the structure and wordplay of their source material, while adding a goofy, appropriately robotic sense of stiltedness. They all represent a new form of comedy, a human-computer collaboration, one that "gathers all these evocative phrases from a genre, and then builds them together in an absurd collage," says Botnik cofounder Jamie Brew.

"Innovation," Jeff Bezos once said, "happens by gently lifting a grandfather and asking him for six different ideas." It's the work of Botnik, a new AI-assisted humor application that scours various types of human-created, word-crowded content--from season-three Seinfeld scripts to Yelp reviews to Bezos' shareholder letters--in order to build predictive, idiom-specific keyboards. Those keyboards, many of which are available on Botnik.org, The best Botnik creations, like this PBS-derived set of otter facts, retain the structure and wordplay of their source material, while adding a goofy, appropriately robotic sense of stiltedness. They all represent a new form of comedy, a human-computer collaboration, one that "gathers all these evocative phrases from a genre, and then builds them together in an absurd collage," says Botnik cofounder Jamie Brew. Botnik began in earnest last year, when Brew--then a writer for The Onion's site Clickhole--began talking with Bob Mankoff, the artist and former New Yorker cartoon editor who, in 2005, launched that magazine's popular caption-writing contest.