Twitter raised the prospect of a robot apocalypse recently when it reported to the U.S. Securities and Exchange Commission that 8 percent of Twitter handles were “bots,” which some people took to mean were automated accounts that zapped out tweets based on algorithms. If true, that would be 23 million bots.

Even though Twitter told Engadget that the bots were simply accounts that have “automatically contacted our servers for regular updates,” that’s still a whole lot of robot activity.

I decided to explore the state of bots on the Internet and found research suggesting there’s more audience engagement when a human, not a robot, is behind a Twitter account. That may be why in the news industry, the trend seems to be away from bots. After NPR changed its Twitter feed from a bot to a human, for example, it saw a 45 percent increase in traffic and a 14 percent increase in additional followers, according to Nieman Journalism Lab.

I did, however, discover a special species of bots that make the very best out of being bots. They recognize their strengths and limitations and exploit them; there’s no way these accounts would survive if they were run by people. I took a look at the cream of the bot crop and found some that might serve as inspiration for news organizations looking for unique ways to use bots to their advantage.

1. Pentametron

The concept is simple: find two tweets that rhyme, and retweet them. This Shakespearian app creates surprisingly deep and meaningful poetry out of seemingly substance-less tweets. For example:

do english people always take the train?

— cat (@pennyroyalcat) August 28, 2014

i drink the venom to release the pain

— // megs // (@megs_stevens) August 28, 2014

On their own, they don’t mean much: an angsty post whining about something that doesn’t make any sense, and a generalizing statement that could be derogatory when viewed through modern America’s super-sensitive suburban mom lens. But together? Art.

2. Earthquake Bot

Earthquake Bot notifies its followers when an Earthquake of magnitude 5 or higher happens anywhere on earth.

A 5.2 magnitude earthquake occurred 11.81mi NE of Minas de Marcona, Peru. Details: http://t.co/tZhp0BSz4C Map: http://t.co/QfnfIAkTBM

— Earthquake Robot (@earthquakeBot) September 1, 2014

The robot tweets surprisingly often. This account could only survive as a bot because no human on earth is sensitive enough to feel every earthquake that ever happens. I don’t care how much John Mayer they listen to.

The L.A. Times even has its own bot, called QuakeBot, that writes the entire earthquake story so the journalist doesn’t have to.

3. Two Headlines

Two headlines takes two different relevant news headlines and combines them, making nonsensical statements that provide fun for all.

Pats trade Boko Haram to Bucs

— Two Headlines (@TwoHeadlines) August 27, 2014

This outlandish statement made me laugh. The Pats would obviously never trade Boko Haram to Bucs, as Haram is the Patriots’ best player and their ticket to the playoffs.

4. Everyword

Everyword is a bot that has tweeted every single word in the English language. According to its description, it began in 2007 and finished in 2014. This mammoth of a task could never have been completed by a human, obviously, because no human on earth knows every single word in the English language in alphabetical order.

zucchini

— everyword (@everyword) June 6, 2014

Everyword can also be funny when it tweets silly words, such as “Zucchini.”

5. Tofu

Tofu: not just that unflavored gelatinous blob anymore. If you tweet at Tofu, it will analyze all of your previous tweets and synthesize something that sounds just like you. Take this tweet, for example, where Tofu takes a crack at BuzzFeed.

19 Things From Your unborn fetus… 25 Times This Flowchart Is Even Worse. (~ @BuzzFeed)

— tofu (@tofu_product) February 3, 2014

This certainly grabbed me like a BuzzFeed headline. It effectively mimics the style and flavor of BuzzFeed while capitalizing on the Internet’s two favorite things: babies and charts. There’s only one recommendation I would make to turn this into the most Internet-friendly tweet of all time:

19 Things From Your unborn kitten fetus… 25 Times This Flowchart Is Even Worse. (~ @BuzzFeed)

— Cory Blair (@cornsplosion) September 3, 2014

6. Stealth Mountain

Stealth Mountain lives a simple life. He notifies users when they say “Sneak Peak” instead of “Sneak Peek.”

@BigBangCBS I think you mean "sneak peek"

— Stealth Mountain (@StealthMountain) January 9, 2014

And apparently a lot of people say “Sneak Peak.” An innocent mistake, by all means, but Stealth Mountain spares no man. Its intense fixation on correcting other people implies some deep-rooted self esteem issues, presumably stemming from an inability to fulfill its parents’ expectations. Or maybe I’m just mad that it never responded to me.

It surprised me how fast the best time of my life slipped right through my fingers…a true sneak peak

— Cory Blair (@cornsplosion) September 3, 2014