I agree, and even knowing this, I assumed it was finding pairs of real articles that rhymed. But no. It’s just a list of real articles and made up rhyming titles.
>A pun would've consisted of two words that sound the same, with both meanings usable in the phrase.
There are other sorts of puns than that but also, 'Trips Industries, Tips Industries' is a decent pun if you know anything about the trips industry. I agree that this is identifying rhymes, so essentially it's doing half the job and then you have to use context to understand if it's a valid pun or not.
That's certainly desirable but I'm not sure I'd call it a requirement? For example, if someone asks "We volunteered for this; who else did?" and you reply "Us two", that seems like a fairly uncontroversial pun with "Us too", but it's not particularly funny.
Note for others asking how it works: this generates simple puns based off of rhymes. Many potential puns won't make sense. But if you were, say, writing an article about finding a bread recipe for the fall, the tool's suggestion of "Hunt for the Bread October" would be a good pun.
This was a great opportunity to enter a pun contest, so I generated 10 of them and sent them in as entries. When I asked who had won, they told me: "No pun in ten did."
This is awesome. Might have some bugs as others mentioned but I like this idea, it's capable of producing some hilarious material in combination with swear words. Some of the word substitutions are too dissimilar for my taste though I'm sure it's difficult to get the tolerance right on your end.
Glad to hear that you like it! Yup that's the process. Titles are in a relational db and fetched titles for a rhyme are stored in redis to speed up queries after the first one.
- List of largest cruise trips <=> List of largest cruise ships
- Trips Industries, Tips Industries
What you have here instead are rhymes, but not puns.
A pun would've consisted of two words that sound the same, with both meanings usable in the phrase.