in 2025, we have a new source for these vague falsehoods: the humble large language model. LLMs can synthesise "new wive's tales" on demand, without the user even needing to ask.
ask for methods to prevent cheese sliding off pizza and be told to add glue. ask for something absurd, like a a conversion of kilometres to oranges, and get a fun story about how many oranges a horse will eat to travel a thousand kilometres. google's summaries even helpfully cite their sources, regardless of whether the source contains the information presented.
folkloric remedies are alive and well in the modern era. we don't need to worry about running out - the biggest tech companies on the planet are pouring billions of dollars and petabytes of pirated knowledge into generating all-new factoids every single day. ask a question to a probabilistic text synthesiser that hates to say no and you're sure to see something fun.