This lab was an interesting one for me not because of the novelty of coding (I’ve been coding for about 7 years on and off) but because of what we were coding: bots. I ended up not getting my bot program working due to being, shall we say, “overly ambitious”, but in the process learned about some other interesting ‘bots in the biz. One person I know, Darius Kazemi, apparently has written a fair bit about twitterbots, and even wrote a Rap-Bot a while back using Wordnik’s rhyming API (which I was tempted to try to learn for my final UnEssay, but decided would eat up too much of my time). I know him through the IGDA Boston Post Mortem game developer meetups, and find the parallel between games, novels, art, and bots an interesting one to explore. Looking through his portfolio of projects, I started thinking of how good a bot has to be before it can pass for a competent, or even good writer or tweeter… and when we’re irrelevant to that equation, then.
Take a look at this tweet, from one of his bots:
Twitter’s Other Space-Time Problem: At 169 Minutes, It’s Too Long to Be a Hit
— Two Headlines (@TwoHeadlines) November 18, 2014
Now take a look at this tweet from CNET:
Woman gets microchip implant so she can leave office keys at home http://t.co/j5aSwqEdNp
— CNET (@CNET) November 19, 2014
Frankly, in 140 characters, it’s hard to tell which is more serious here. The fact that a major tech news outlet like CNET crafts tweets as strange-sounding as a bot’s mashing up of headlines like that is a bit disconcerting. Someone took the time to write that second tweet; hell, an SEO team probably got paid a lot of money to agonize over the tweet’s viral potential, its word choices, and more. I’ve written for a tech outlet, I know it’s a thing they do.
To be clear, I’m not saying this is an example of Turing-test-passing bots; I’m saying that they’re getting to a point where they’re gumming up Twitter with enough realism that the site’s eventually going to need to require programmers to label their bots as such (or have a way to get them tagged), or we could end up with some crazy situations of people taking a headline out of context (because let’s be honest, who on Twitter hasn’t seen a headline, retweeted it, and then regretted it once they read the article? Anyone? No? Liars, the lot of you) and that could be messy.
Basically, this lab just got me thinking about a lot of silly things that may or may not come to pass, and while they may not be of much importance to me personally, I can see where the communication breakdowns can occur if bot programming gets more advanced. What I cannot see is where the error in my Python code is…
In other news, we just missed Bot Summit 2014 on November 8th, in Boston, and I’m upset I didn’t know about this sooner. There’s a video of it on Darius’ website, linked above.