I’m one of the 150K followers of the Twitter bot @horse_ebooks. I just built my first Twitter bot, @horse_edubooks, based upon it.

Why follow a bot? Why build one?

No doubt, there are lots of bots on Twitter that simply spew about weight-loss drugs or auto-respond to keywords like iPad. With their link-bait tweets, it’s easy to dismiss these bots as spam. But there are some creative bots on Twitter too. @pentametron, for example, finds and retweets rhyming couplets in iambic pentameter. Robot J. McCarthy (@RedScareB0t) responds sternly when someone mentions “communism” or “Marxism.”

And then there’s @horse_ebooks, which tweets out random scraps of text (well, along with links to an online e-bookstore). A bit spam. A bit nonsense. A bit poetry.

Demand Any Day Of The Week!

— Horse ebooks (@Horse_ebooks) January 20, 2013

My Twitter bot, @horse_edubooks, is based on this and built using the code that Boone Gorges wrote for @horse_thatbooks. That source code is available on Github.

The bot pulls the tweets from a hashtag (#edchat in the case of the former; #thatcamp, the latter), stores them in a file, runs these through a Markov chain process, and then tweets out the result.

Insight from peers & learn until they get connected. #edchat

— horse_edubooks (@Horse_edubooks) January 21, 2013

Dave Lester recently made a great case for “why we should be discussing bots”: the automation of mundane or time-intensive tasks; the “monitoring, notification, and extending [of] oneself through software”; the moving of "agency to the edges of an ecosystem"; modularity; interoperability; problem-based learning; fun.