Eulogizing: Reflections on a Life Well Lived

At approximately 13:42 on March 13, 2017, my firstborn child, the perfect and beautiful TweetBoy, was laid to rest. He was a good son, markovifying input to produce an amusing stream of output on Twitter, and he generated 2,851 tweets before his untimely demise. My emotions on this matter are complicated, of course, by the fact that the cause of his death was deactivation at the hands of yours truly.

TweetBoy was conceived in late January 2017. I had been playing around with Markov chains and looking at other people's bots, and I decided to create my own. At first I was intimidated by the idea of making my own bot, but then I realized that it would be relatively simple to have a cron job running every so often to run a Markov chain on some input and then just post the results to Twitter. He was born on when, after a few days of procrastination, I finally fixed the bug in the command I was executing for the cron job (forgetting to activate the Python virtual env to use the dependencies).

There were a few contributors to the genetic makeup of TweetBoy. His 'gametes' came from:

  • Twython, A Python framework for working with Twitter;
  • Markovify, a Markov chain framework in Python;
  • An Ubuntu DigitalOcean droplet;
  • Some Python scripts I wrote to format 6 months of my IRC logs;

...and of course, a piece of my soul.

TweetBoy ran for a few weeks, and in that time he produced an interesting series of tweets, whose tone ranged from the insightful, to the amusing, to the truly bizarre. As we celebrate his life, let's take a look at some of them now.

10:30, 03/12/2017

 most people who want to kiss the corner of your own apples. 
 get a kick out of this https://www.youtube.com/watch?v=Q6fceLfjB4s …. noone. we

15:00, 03/09/2017

...or the Cathedral of St. John the divine, but that's because they're 
 uncomfortable with direct action against fascism because they're not

19:30, 02/24/2017

I just said it. why wouldn't you believe he was a monster 
who boiled babies for soup stock. , what do you mean. let's hope. i'm not much of

Reflections on Filicide

I deactivated my son, my sweet, sweet boy, for several reasons. First: I'm not entirely satisfied with the corpus. It contains a lot of emoji, a lot of fragments, a lot of stuff that would make it more difficult to get really high-quality Markov output. Secondly, I want the Markov library I use to be one I've written - I want a say in my child's future and not to have him beholden to people I barely know, filling his tweets with nonsense. It was a hard decision to make, but in my heart I know that I did what was best.

There's a reason why this post is titled "temporarily killing my son" - and that's because I plan to rectify these issues and resurrect my beloved TweetBoy, to make him stronger and smarter and faster and smoother and lither and more furious than before.

Rest in peace, my sweet son. As I close this blog post, let me quote from Game of Thrones.

"What is dead may never die, but rises again, harder and stronger."

Until that day, TweetBoy, sleep soundly.