Some weird text-generation stuff I’ve enjoyed recently

Since working on Verified Facts, I’ve been paying a lot more attention to web projects that do similar things with text and text randomization. Here are a few projects that have entertained me lately.

Upworthy Headline Generator

upworthygeneratorAs we learned with Verified Facts, making a parody using randomized text is a great way to point out how meaningless, repetitive, or shallow something is. “This is so bad… it can be copied by brainless text robots!” is a pretty harsh burn. I’ve seen and enjoyed my fair share of Upworthy videos, but the over-the-top, teasing, and occasionally-exploitative titles can get pretty damn annoying. This Upworthy generator totally nails the textual style of those headlines, while the “youtube images” chosen for the body of the page also nail Upworthy’s tone and subject-matter.

I don’t really have a problem with Upworthy, but I admired the “accuracy” of this particular generator a lot.

What Would I Say?


What Would I Say? combs your Facebook posts and comments and comes up with a randomized utterance made up of various parts of sentences you’ve previously spoken. It’s amusing but a little repetitive. Here’s a full album of things it’s predicted I would say.

Headline Smasher


This one’s excellent. It allows users to draw sentence fragments from a wide range of news and opinion websites and to generate crazy-sounding article headlines. The cool thing is that when a user is generating headlines, they can select which sources they would like the program to draw from.

One of the biggest ways that users interact with any random-text-generation tool is through selecting or curating the results they would like to share with their friends. This goes for web toys, twitter bots– everything with randomization invites human curators. Most random text generators produce a mix of both excellent and boring output. When we made Verified Facts, we knew that it would be important to let users save results and share the ones that made them laugh the most.

Headline Smasher’s source selection improves on this by allowing users to an additional degree of curation. If I’m in the mood for, say, headlines about celebrities, I can choose to mix celebrity news sites with specific other types of sites. The results are still random, but I’m providing greater input, which allows me to see only the content that will make the laugh the most. And the fact that the funny results were based on my advice also strokes my ego, and makes me feel cleverer than if I were just an observer.