#barrl.net

[norareedbarrl]

some word lists for bot-making

After a couple years of making bots, I’ve made a fair few wordlists that I thought other people might be able to make use of, so I decided to release them. These are provided under a Creative Commons Attribution-NonCommercial 4.0 International License; if you want to use them commercially, please contact me and we can work something out.

I’m just providing these as lists in txt files; I use delim.co to turn them into Tracery code.

body parts

goo

buildings that are full of secrets

unsettling gerunds

unsettling adjectives

cooking tools

ingredients

cooking styles

dishes (compiled from multiple cookbooks, mainly the Joy of Cooking, as well as suggestions from Twitter followers)

  some word lists for bot-making was originally published on barrl

[norareedbarrl]

a big list of bot-making resources

I’ve been talking about how easy it is to get started making bots for a while now, but I haven’t done a long-form writeup of all of the tools that I use that you might find useful. So here we go: a list of resources that I use for making Twitter bots.

Tracery Basics

Cheap Bots Done Quick This is the software that most of my Twitter bots run on. It handles the regularly scheduled posting as well as the replies

Tracery The language my bots are written in

Tracery Visual Editor

Another Tracery Editor

Tracery Interactive Tutorial

Non-Tracery Tools

Auto-fave/Retweet bots I use these for the honeybots

Zach Whalen’s SSBot

Word Lists

SCOWL Wordlist creator

Ashley Bovan’s Word Lists Includes words by syllable count, hyphenated words, words by parts of speech, and other lists

The Great Noun List

Corpora Lists of interesting words

List inspiration

The Noun Project I never use any of the Noun Project stuff directly, but they have a lot of lists of icons of various objects put in useful categories, which can be good for when you need to think about, say, common appliances, but it’s hard to name them all off the top of your head

Spellzone word lists These are for learning English, but can also be useful for when you need a bunch of words for certain things, like buildings, appliances, household objects, etc; I usually just keep a notepad tab open and write down a handful that will work for what I’m doing

Wordnik I often will pick a fun sounding word, look at what lists people put it on, and then look at those lists for more word inspiration

Text/List Editing Tools

Whitespace remover

Comma Delimiter Turns lists of words into code; use commas as separator and double quotes for Tracery

Find and replace In case the one in wordpad/notepad isn’t quite enough

Case Converter

Dolch Sight Words This is actually a word list for teaching kids to read fluently, but it can work well for basic bot vocabulary

Images/Graphics

Public Domain Photos

Vintage Printables This one’s particularly good for technical illustrations

Wikimedia Commons

Glitch’s game assets, PNG conversions

Miscellaneous Guides

Instant disposable Gmail addresses for signing up for Twitter

Hook, bait and camouflage: making a honeybot

Further Reading

Ethical bot-making An overview of bot-ethics with lots of good links

So you want to build a generator…

a big list of bot-making resources was originally published on barrl

[norareedbarrl]

What's with @thinkpiecebot lately?

It’s been over a year since I started @thinkpiecebot, and it’s grown a LOT in the interim. The code is a huge, often unwieldy mess, there are over 100 formulas in it, and in the past couple of months the scope and the tone has gone through some… changes.

Holy Shit, a Suit-Wearing Racist

— Thinkpiece Bot (@thinkpiecebot) November 21, 2016

It was a rough election year, and since we elected a white supremacist sexual predator as the next president of the United States, it’s only gotten worse.  We’re watching a fascist gather military leaders around him and fuck up foreign policy for his own profit. Abortion bans gaining traction, hate crimes going up, reasons to crack down on the poor, on POC, on trans people, all going into overdrive.

It wasn’t good to begin with. At any given time, I know about a dozen people crowdfunding their medical expenses, and more people their other survival needs.

Meanwhile, the discourse has been absolute garbage. We keep seeing fawning profiles and photo shoots with the ringleaders of hate groups, normalizing them. A bunch of dudes on the left have managed to normalize a political environment that mostly consists of them taking screenshots of marginalized leftists on Twitter and making jokes about and/or harassing them; to add insult to injury, the jokes aren’t even very good. This happened. Political propaganda and coordinated disinformation campaigns are being called “fake news” and confused with The Onion, which also isn’t very good anymore. Dudes keep calling for “civility” and “respect” for the people who supported our fascist president-elect. The Daily Show gave a white supremacist a platform that got her a NY Times puff piece, then sent her cupcakes. That’s just a sample.

Personally, I’ve been dealing with a group of stalkers that a prominent “feminist activist” sent on me. I have various groups of people, some of them trans, showing up to tell me my gender isn’t real at least once a week now. Throw in some medical problems and mostly I’m mostly a ball of exhaustion and low-simmering rage that happens to take human form and make Twitter bots, and if I look at my friends, I see that I’ve still got it pretty good.

When I started thinkpiecebot, I was mostly making fun of terrible op-eds by baby boomers where they pretend they didn’t ruin the economy and fuck over our entire generation. Lately, though, I’ve been pretty much channeling all of my frustration and rage at the current state of political discourse, the utter failures of journalism to cover what this is– a full-scale fascist revival, and the mostly cishet white dudes with platforms who refuse to step aside and let people who aren’t so privilege-blinded that they don’t know what they’re talking about do their jobs. The total spinelessness of both the liberal media and the Democratic party has been predictable but disappointing, and I’m seeing a lot of the people I care about be thrown under the bus in an attempt to be conciliatory with neo-nazis. And, well, this bot is the platform I have, and the headlines are still terrible enough that I can riff on that.

Thinkpiecebot is built in Tracery, and it works by having a series of formulas with lists of things that get plugged in. For example, “[EXCLAMATION], A [MODIFIER] [BIGOT TYPE]” can make tweets like “Holy Shit, A Suit-Wearing Racist”. Most of my formulas have pretty long lists of words that can get plugged into them, but I’m at a point where I have over 100 formulas, so I can put in a few that have a more limited selection of possible phrases and not worry about them naturally coming up more than every few days. I’ve been putting in a lot more formulas that are extremely pointed criticism at specific types of stories.  

I Wrote Milquetoast Liberal Thinkpieces. Now I’m Collaborating With Nazis. — Thinkpiece Bot (@thinkpiecebot) December 1, 2016

This isn’t an accident, and it isn’t a bot that learns automatically picking something up. This is me, attempting to use satire to point out something that the media is utterly failing at, assisted by a random number generator. Sometimes it works really well. For example:

Well, The United States Electing A Sexual Predator As President Is Normal Now, Let’s Talk About Cupcake Shops Instead

— Thinkpiece Bot (@thinkpiecebot) December 7, 2016

What’s great about the evolution of thinkpiecebot is that a lot of the pieces I was originally mocking were about various forms of completely trivial bullshit, from cupcake shops to comic books, and I’ve been expanding it for over a year. That’s a year of reading terrible thinkpieces, taking suggestions from people on Twitter, and sending drunk emails to myself while my friends are in the bathroom to remind myself to put the stuff we were talking about in when I get home. It means there’s over 1,500 unique phrases in the bot that get plugged into the 100+ formulas.

I can’t hold all the stuff that goes in there in my head at a time; I’d never have thought to put “Cupcake Shops” in the above tweet. That’s the power of the random number generator. When you combine it with phrases like “The United States Electing A Sexual Predator As President” and “Headlines That Are Too Fucking Precious To Use The Word ‘Fascism'” that are obviously written by me in my voice, the bot starts to sound less like a bot and more like me.

The bot’s less funny now. But sticking with making fun of the kind of shitty thinkpiece that was bugging me in the middle of 2015 would have just made it slip into irrelevance, and from the beginning, thinkpiecebot has been a living project that expands and changes. I’m longing for a time when the thinkpieces that made me mad were just boomer nonsense about Pokémon, but I think this bot is gonna keep getting darker for a while ago. Maybe it will be funnier if we can fix the world.

What’s with @thinkpiecebot lately? was originally published on barrl

[norareedbarrl]

the saga of @christianmom18

Over the past few months, I’ve been making a handful of “honeybots”– bots that act as a honeypot for Twitter troll. There are a lot of people on Twitter who search for specific terms and then yell at people who mention them; they go on about topics that range from chemtrails and the flat earth to various alt-right people with cult followings to atheism. A handful of bullies particularly like to search for people who mention them negatively and then retweet those people to their followers, leading a harassment mob to their virtual door.

I’ve been documenting these on a Tumblr blog, and they caught the attention of the Washington Post, among other places. Sarah Nyberg made an advanced, improved honeybot that is even better at catching trolls and wasting their time than mine. It’s been a fun couple weeks for automatically wasting the time of assholes.

One of the groups of people who have been really easy to bait and really plentiful are “internet atheists”. If you’ve spent much time adjacent to the atheist community, you’ve probably run into these– their identity revolves around their atheism; they generally hate religion of all stripes, and they roam the wilds of the web trying to pick fights with people. They tend to have a personal philosophy that revolves around the idea of skepticism, but aren’t very good at embracing its principles themselves– they tend to have “scientific” justifications for all of their personal biases and often express a good deal of transphobia, racism and misogyny.

What these people seem to like more than anything else is to fight their idea of “theists” without actually having to engage in ideas. They want a straw Christian to use as a punching bag. So I made them one.

Enter @christianmom18, AKA “carol”.

Carol is a bit of an improvement on @good_opinions and @opinions_good. She has a lot more basic things to say as “bait”, and I made her reply occasionally with those instead of just with the 17 “argument” phrases– “you are wrong”, “check the bible”, “no”, etc. I also gave her a chance of “correcting” with “*your” anytime someone said “you’re” and vice versa, regardless of whether or not they were using the word correctly. (Another great Sarah idea.) I also set her up to auto-retweet some Bible quotes accounts and the official accounts of both Kellogg’s cereals and the New York Yankees.

I got some help from Sarah and some of my gaming friends to come up with this stuff and had “carol” stick emoji and random punctuation on the end of the phrases so that they didn’t get caught by Twitter as duplicates, and practically immediately started getting bites.

(I actually removed the “going to hell” thing from carol’s vocabulary after this– I realized that was kind of an asshole thing to say, even if the only people who saw it were jerks.)

A lot of these accounts are people who spend a lot of time searching Twitter for terms like “atheists” to find people to dunk on, and they often follow each other, so “carol”‘s posts quickly spread among that network, and people started talking to her.

At length.

That’s just the beginning. Richy went on to talk to “her” for about three hours— through several repetitions of her “arguments”.

Some people figured out that she was a troll, but not that she was a bot.

Carol’s maturity was called into question.

She was told she was a bad mother.

  She was insulted:

And this is just the beginning. After this, “carol” was quote tweeted by a popular atheist account that seems mostly to spend time dunking on theists, and her followers came in droves.

Check out how smug these people are:

She was ‘splained to pretty much endlessly:

After about a day of “Atheist Girl”‘s followers talking to carol– a few of them figured out that “she” is a bot, but not many– I started noticing that they were even taking the really obvious bait.

By this point, Twitter’s automated system had shadowbanned “carol”, so her posts weren’t showing up in Twitter search results– all the people talking to her found her via other people’s retweets, quote tweets and conversations. So I got some help from friends to come up with really wacky things to say.

This is just the tip of the iceberg: search Twitter to see all “her” mentions. I also shared her source code, though it’s not free for reuse without permission.

the saga of @christianmom18 was originally published on barrl

[norareedbarrl]

on tessera(e)

  By Urban~commonswiki; used under a CC licence.

  I wrote a piece recently for BotWatch that touched on this, but I wanted to give a brief explanation of why I use the word “tessera” when I talk about my bots.

A tessera is a single tile used in a mosaic; every tessera is suspended in mortar, and depending on where the mosaic’s viewer is standing and the lighting conditions, different tiles may be illuminated. Guy Gavriel Kay explains in Sailing to Sarantium:

“The curve and the height of a dome allow us the illusion of movement through changing light, my lord. Opportunities beyond price. It is the mosaicist’s natural place. His… haven. A painted fresco on a flat wall can do all a mosaic can, and-though many in my guild would call this heresy-it can do more at times. Nothing on Jad’s earth can do what a mosaicist can do on a dome if he sets the tesserae directly on the surface.”

Some of @thinkpiecebot’s tesserae.

My bots have lots of words and phrases in them and they’re set in formulas, what the bot says is chosen pseudorandomly, just like the flicker of a candle may psuedorandomly “choose” different tesserae to illuminate, and how their flickering and winking will be different for people standing in different places.

In some ways, I’m creating the illusion of human thought where there isn’t any; which of my tesserae is “illuminated” depends on what the random number picker that puts bot tweets together happens to land on. Just like changing light creates the illusion of movement, the random numbers create the illusion of a mind behind what the bots post.

on tessera(e) was originally published on barrl

[norareedbarrl]

how and why @nerdgarbagebot works

As usual, I’ve made a few new Twitter bots since my last update, but I want to talk a little bit about how a particular one– @nerdgarbagebot— works.

I’ve been talking a lot on Twitter about how simple bots can be to set up. I use Cheap Bots Done Quick and the Tracery visual editor for mine, so I hardly have to look at code when I’m writing them– my screen actually looks like this:

Both of these are free tools that you can use to make your own bot; all you need to do is register a new Twitter account to get started.

The results of nerdgarbagebot are pretty varied and rich, but the code is really simple– it only actually has 13 variables in it:

Work titles (“Jurassic Park”, “The Sims”)

Creators (“George Lucas”, “Nintendo”)

Elements (“mermaids”, “a plot”, “interpersonal drama”)

Comparative adjectives (“grittier”, “with swears”, etc)

Formats/mediums (“tv show”, “installation art piece”, “webcomic”)

Genres (“gothic”, “cyberpunk”)

Settings (“on the high seas”, “in a modern high school”)

Story types (“coming of age story”, “mystery”)

Characters/People (“Yoda”, “Joseph Gordon-Levitt”)

Character roles (“mentor”, “president”)

Audiences (“tweens”, “atheists”)

“Imagine this” intro phrases (“Imagine”, “I need you to picture”)

“Fund this” intro phrases (“Crowdfund this”, “Please support my”)

For each tweet, Tracery at random chooses from a list of formulas, which are written like this:

It’s a #genre# version of #titles#, with #element#.

#imagine# a #format# version of #titles#.

#imagine# a #genre# #format# version of #titles#, but with #character# as the #role#.

I’m increasingly realizing that the reason bots like this work really well– compared to ones like @BuzzFiendNews, which I am still struggling with improving– is that even though the format is simple, every phrase is something that the bot’s audience brings their own baggage and associations to, so the tweets tend to be more unique in and of themselves.

BuzzFiend is a lot harder to add variation to, because even though each tweet has different words in it, every tweet using a specific formula tends to be similar to others. A joke about eating humans tends to be pretty similar to other jokes about eating humans, whether those humans happen to be sidekicks or princes.

Thinkpiecebot, nerdgarbagebot and some of my other bots, especially @likeuberbut, manage to capitalize on people’s existing ideas. Thinkpiecebot’s funniness in particular comes from the unexpected combinations that it produces being put into the recognizable headline format, but doing that ended up being complex– I have over 50 formulas in it and nearly as many variables, and I’m constantly updating it so that it keeps up with the zeitgeist.

Nerdgarbagebot is brand new, and I’ll probably have to continue updating it to keep things current, but I’m having a lot of fun with it– I hope you like the results too!

how and why @nerdgarbagebot works was originally published on barrl

[norareedbarrl]

4 new bots

In case you don’t follow me on Twitter, I’ve made 4 new bots over there since I last updated: @roboromance, @full_of_eels, @haters_improved and @restartthevoid, which has an artist’s statement. You can see all my bots from the Twitter list; they’re also listed on my projects page.

4 new bots was originally published on barrl

[norareedbarrl]

Make your own @hydratebot: a tutorial for non-coders

So @hydratebot has become pretty popular, and people keep requesting different frequencies of tweets, having themselves @-ed in the tweets, etc, and I’ve been looking for a way to show y’all how easy Cheap Bots Done Quick is so you can set up your own bots. So I’m gonna kill two birds with one stone here and show you how to make your own @hydratebot, which you can modify as you want; just credit @NoraReed in the bot profile.

The first step is creating a new Twitter account. Log into it, then go to Cheap Bots Done Quick and click on the “Sign In With Twitter” button.

You’ll need to authorize the app to post from that account.

You’ll get a big text box. Copy and paste the following into it:

{ “please”: [ “please”, “i command you to”, “you must”, “go now and” ], “have”: [ “have some”, “drink some”, “consume” ], “water”: [ “some water”, “water”, “fluids”, “liquid” ], “address”: [ “human”, “friend human”, “human friend”, “hu-man”, “meat-creature”, “meat-friend”, “squishy human friend”, “biofriend”, “biological being”, “my friend”, “my hu-man friend”, “physical entity” ], “thanks”: [ “thank you”, “thank you #address#”, “thanks”, “thanks #address#”, “thank you from a robot who loves you” ], “hydration”: [ “become hydrated”, “put water into your mouth-hole”, “consume this \”water\” that humans require to live”, “drink hydration”, “put water into your body so that it will function”, “drink water so you can maintain your physical form”, “put liquids into your mouth-hole” ], “timeto”: [ “it is time to”, “you must” ], “standardorigin”: [ “#please# #have# #water#, #address#!”, “#address#! #please# #have# #water#. #thanks#.”, “#address#! #timeto# #hydration#.”, “#timeto# #hydration#, #thanks#.”, “#address#, #please# #have# #water#.” ], “origin”: [ “#standardorigin#” ] }

Now, set it to the frequency you want and click “save”.

Want it to @ you? Scroll down to the bottom, where it says “origin”. Insert your Twitter handle before #standardorigin#. This is handy if you want it to take advantage of Twitter’s @ notifications.

More Advanced Stuff

If you want to change it in other ways, Tracery’s visual editor is pretty easy to use; just click on “JSON” and copy and paste the code there, then click “JSON” again to get an easy interface:

GalaxyKate has Tracery tutorial if you want to do anything advanced with it.

This particular bot has only a single statement for #origin# so that it’s easy to add handles to it without having to modify every #origin# phrase.

You can also use IFTTT to easily connect your tweets to hundreds of other services.

That’s it! Feel free to contact me on Twitter if you have any questions. If you found this useful, consider buying me a coffee or supporting my Patreon. Thanks! Happy hydration!

Make your own @hydratebot: a tutorial for non-coders was originally published on barrl