Porter Olsen: It’s my great plea­sure to intro­duce Darius Kazemi. He began his career as a game devel­op­er, which he did for ten years and worked on mas­sive­ly mul­ti­play­er online games includ­ing Lord of the Rings Online, which I played a great deal. But of course if it said Tolkien I was inter­est­ed to at least check it out. Then he decid­ed to write an arti­cle titled ​“Fuck Video Games” and that was the end of that.

But he pur­sued his career as an author. He’s recent­ly released a book on Jagged Alliance 2, a turn-based squad strat­e­gy game. He’s per­haps most well-known for his work in the ​“weird Internet,” specif­i­cal­ly bot-making and a pro­gram that he wrote called Random Shopper, which goes to Amazon and sys­tem­at­i­cal­ly spends fifty dol­lars a month buy­ing him stuff which you actu­al­ly buy, right? [To Darius] And then it gets shipped to you and you see what you get.

He also writes Twitter bots includ­ing a muse­um bot that goes to a col­lec­tion and ran­dom­ly shows you some of the con­tent that is not nor­mal­ly seen in the nor­mal course of vis­it­ing the muse­um. I’m very much look­ing for­ward to the time at which Random Shopper buys for you Jagged Alliance the book. Thank you for being here. We’re excit­ed to hear your talk, and I’ll turn it over to you.

Darius Kazemi: Hi, every­body, I’m Darius Kazemi and I hope I can live up to the abstract that I sent in ever so long ago. I do want this to be a dia­logue and not just me up here so I do have a talk pre­pared. It it not going to fill a full hour because I expect to talk to you all while we’re here. 

That was a pret­ty good intro­duc­tion. As Porter men­tioned I spent the first part of my career in video games. I’m actu­al­ly going to talk about that a lit­tle bit because it relates to what I want to dis­cuss today. 

“Stuff I made in 2013″ at Darius’ blog

I make a lot of stuff. This is the projects that I released in 2013. Actually a few more than that. This is like 18 months of projects, and each one of these is—my book that took me a year to write is one entry on here, and lit­tle projects that took me an hour to release are also list­ed on here. I con­sid­er them all of equal val­ue because I firm­ly believe that there is almost no cor­re­la­tion between how much work into some­thing and whether or not it will make an impact, peo­ple will like it, that sort of thing. All I know is that I like all this stuff.

I’m here at MITH today, and I want­ed to talk a lit­tle bit about dig­i­tal human­i­ties from my posi­tion as an inter­est­ed out­sider. I’ve always kept a fin­ger in acad­e­mia, at first through game stud­ies and peo­ple study­ing video games, and more recent­ly through elec­tron­ic lit­er­a­ture and those fields. I’m not going to go into a ​“what is it?” debate because I know every­one who’s in dig­i­tal human­i­ties is very tired of those, but we know when we see it, right? But I do want to talk about what an out­sider thinks dig­i­tal human­i­ties is, which is prob­a­bly best encap­su­lat­ed by the Ngram Viewer. It’s this sort of Big Data approach to the human­i­ties that has been writ­ten about a lot, and peo­ple have writ­ten about why it’s bad from a cul­tur­al and insti­tu­tion­al cri­tique and those stand­points, and how it’s tied to Silicon Valley trends, tied up with the cor­po­ra­ti­za­tion of the uni­ver­si­ty, and so on and so forth. I’m not that inter­est­ed in cov­er­ing those angles, but what I would like to bring to this for a bit is my own expe­ri­ence using these sorts of ana­lyt­i­cal tools onto the at least some­what human­is­tic enter­prise of the video game.

I spent five years work­ing as a data ana­lyst for online worlds. I start­ed work­ing at Turbine in Boston. That was in 2005 and we were just begin­ning to under­stand what we had on our hands in terms of data. When I start­ed work­ing there on Dungeons & Dragons Online, we had a serv­er that was col­lect­ing data about what our play­ers did on the servers, but we had no report­ing on it. No one was using it. They were just these data­bas­es that got slow­ly filled over time, until occa­sion­al­ly some admin would go and write a script that would blow away any­thing over six months. That data was­n’t even engi­neered for the pur­pose of track­ing play­er behav­ior. It was most­ly engi­neered by engi­neers who in the course of their bug-fixing decid­ed it would be help­ful to write a track­er to do this, or a track­er to do that. By the time the game was launched, we had all these track­ers that were col­lect­ing data that was meant to be tech­ni­cal, but could also be ana­lyzed to learn things about the way that play­ers acted. 

But no one was look­ing at this infor­ma­tion at all. I learned about this serv­er and I was this enter­pris­ing 22 year-old with an engi­neer­ing degree and a minor in rhetoric, so I thought I could do any­thing. I fig­ured that I could build soft­ware that turned that data into dash­boards of infor­ma­tion that we could then act on. So I did that. I was work­ing as a tester at the time, but I got per­mis­sion to do this project on the basis that it would save us mon­ey and time test­ing if we could out­source a lot of our test­ing to the play­ers them­selves. We could find out that peo­ple are wip­ing out in a par­tic­u­lar dun­geon if we get sta­tis­tics on whether or not peo­ple are suc­ceed­ing in a dungeon.

This was a huge suc­cess inter­nal­ly at Turbine. In an 18 month span, we went from no for­mal report­ing appa­ra­tus at all to a team of four peo­ple ded­i­cat­ed almost exclu­sive­ly to col­lec­tion and inter­pre­ta­tion of data, with myself as the chief inter­preter. I like to say that we cre­at­ed this cul­ture of data at Turbine. We went from a com­pa­ny that did­n’t talk about these things to a com­pa­ny where design­ers would come to me before they even had imple­ment­ed a new idea that they had, and said, ​“I have these con­cerns. What can we do to track data about our play­ers that can then tell me once this fea­ture goes live whether or not it was a success?” 

And this was new in the indus­try at the time, as well. I believe there was pret­ty much one oth­er com­pa­ny that was doing data analy­sis at this lev­el, and that was EA on The Sims Online. Most of the oth­er com­pa­nies were sim­ply scrap­ing log files and occa­sion­al­ly look­ing at num­bers here and there. But at best a com­pa­ny would have a sin­gle engi­neer who would do this part-time because it seemed like a good idea. So this was all very new. This was about four years before Zynga real­ly pop­u­lar­ized social online games and col­lect­ing met­rics and tun­ing their game based on that, to the point where I had to go out to con­fer­ences and try and con­vince peo­ple in video games that it might be a good idea to col­lect data on their users.

Virtual worlds are real­ly inter­est­ing for this kind of work because in a vir­tu­al world, you’re pre­sent­ed with what seems like this closed epis­te­mo­log­i­cal sys­tem, where in the­o­ry if it’s hap­pen­ing in the sim­u­la­tion, you tech­ni­cal­ly have per­fect data on what’s hap­pen­ing with­in the para­me­ters of that sim­u­la­tion. It seemed like we were solv­ing epis­te­mo­log­i­cal prob­lems inher­ent to user stud­ies and play-testing. We no longer had to deal with all that messy ​“Well are they real­ly play­ing like they would if we’d brought them into the com­pa­ny? Are they say­ing nicer things about the game because we bought them piz­za?” All the way down to hairi­er prob­lems than that as well. So it seemed like we were doing this very sci­en­tif­ic thing.

What I’ve learned from that expe­ri­ence is that using data can get you very far…until it does­n’t any­more. One of the hard­est lessons that I learned, and I learned it slow­ly over the course of three years, is that you’re almost always con­firm­ing sus­pi­cions that you already had. It’s unclear whether you’re com­ing in ask­ing biased ques­tions, and in fact even the very nature of the sta­tis­tics them­selves can be called into ques­tion as to whether or not they’re accu­rate or mean anything.

We built all sorts of visu­al­iza­tions. This was from my con­sult­ing com­pa­ny that I had after that. We looked at peo­ple play­ing Quake and built these beau­ti­ful visu­al­iza­tions of how they did it, and you could look at that and make con­jec­tures about it. Next up is an exam­ple that I don’t think I’ve shown before. This is a cap­ture from a tool that I built inter­nal­ly at Turbine. This isn’t even live data, this is from our beta serv­er. I built this tool called the Economy Explorer, and we used it to look at trade net­works in the game and sort of fer­ret out who the gold farm­ers were, and who’s doing real­ly strange things trad­ing with a bunch of place­hold­er char­ac­ters and then fig­ur­ing out who the mule is and cut­ting them out of the net­work, and all that kind of stuff. 

[Here Darius plays a clip ~11:12–11:55 in the video show­ing the application.]

It was very excit­ing work, but this was the build­ing of a tool to help an author­i­tar­i­an appa­ra­tus in the game, essen­tial­ly. This was the tool that we used to ban peo­ple for doing things that we did­n’t like in the game.

When it game down to actu­al insight and fig­ur­ing out fun­da­men­tal things about our game and learn­ing about it, I used to get this ques­tion at con­fer­ences all the time and I always hat­ed it, which is, ​“What’s the most sur­pris­ing insight you uncov­ered through your data?” I nev­er had a good answer for this ques­tion. Usually I’d default to some­thing like ​“Well the big sur­prise was that there were no big surprises.”

And yeah, we’d uncov­er unusu­al stuff but it usu­al­ly just mys­ti­fied us. We were smart peo­ple, I had a strong human­i­ties back­ground in addi­tion to a tech back­ground. I knew things, I had cross-disciplinary knowl­edge, like I thought I was smart enough to slice and dice it and fig­ure out what this stuff meant, but the more I sliced and diced it, the more it was like infi­nite regres­sion of nest­ing dolls. You just open up this box and then there was anoth­er box, and then you open up that box and anoth­er box. It’s prob­a­bly a human head at the very bottom.

We’d uncov­er things like a whole bunch of peo­ple are buy­ing this par­tic­u­lar kind of item. Sometimes we’d look at it and at best we would notice that the item’s pret­ty over-powered for the price. Well great, that’s use­ful, but it’s not sur­pris­ing. It’s not an insight that our play­ers are opti­miz­ing for the sys­tems that we told them to opti­mize for. Over and over and over again, it just told us ​“things are going the way you prob­a­bly expect them to go, but maybe they’re going worse or maybe they’re going a lit­tle bet­ter” and that was pret­ty much it.

People would want to know about mon­ey and pur­chas­ing habits, and that’s what we actu­al­ly got out of the realm of the epis­te­mo­log­i­cal shield that we had set up because now you’re deal­ing with humans with cred­it cards, and there were pri­va­cy laws and all these oth­er things. So I sort of said okay, I’ll do the busi­ness ana­lyt­ics on that side because I have to for the busi­ness peo­ple, but on this oth­er side is the game stuff look­ing at the game as a work on its own mer­it, and I can still slice and dice that and learn things. Over and over again we either learned that our sus­pi­cions were found­ed, which either meant that our sus­pi­cions were always right or that the way we were col­lect­ing this data was just set up to con­firm our sus­pi­cions, or we found weird things. For exam­ple, often­times when we’d find some­thing weird it was like this sto­ry I’m about to tell.

So we have mon­sters run­ning around this world, because of course you’re not going to have a beau­ti­ful vir­tu­al world with­out mon­sters get­ting in your way all the time. We put in a few hooks to the arti­fi­cial intel­li­gence of the mon­sters because some of the engi­neers were wor­ried about the pathfind­ing for the mon­sters; that they would get stuck in cer­tain places. So essen­tial­ly a mon­ster would walk around, and at any point in its pro­gram if it did­n’t know what to do next and said ​“I lost my march­ing orders. I’m com­plete­ly lost. I have no idea what to do.” it would ping out those coor­di­nates and it would just say ​“Okay, I got con­fused here.”

So said I could make a map of all these places the mon­sters get con­fused and maybe I could find some weird place where the geom­e­try is bad and needs to be fixed in the game. So we pull up the map, and the map looks almost like an exact grid, like all these dots but the dots are all clus­tered along these lines mak­ing this grid. I had no idea. I thought there was a bug in my code that col­lect­ed this data or some­thing. I took it to a much more expe­ri­enced engi­neer who took a look at it and said, ​“I know exact­ly what that is.” Our sys­tem is divid­ed into land blocks, and the AI just gets con­fused when it’s basi­cal­ly cross­ing a latitude/longitude line into anoth­er par­cel of the serv­er. It gets momen­tar­i­ly con­fused and then reori­ents itself. Invisible to a play­er, but got sent out to the server.

So that’s weird and sort of inter­est­ing, but it reminds me of this. Anyone who’s ever searched for ​“fuck” on the his­tor­i­cal ngrams site, you prob­a­bly were mild­ly sur­prised to note that there’s a lot of f‑bombs being dropped in 1650, until you go to Google Books and actu­al­ly search for that word in that time peri­od. Then it turns out that it’s just the word ​“suck” or some­times ​“such” depend­ing on the font, and the OCR picks it up the wrong way. That reminds me of find­ing that weird grid pat­tern with those arti­fi­cial intel­li­gence enti­ties in the game. It’s equiv­a­lent to those AI seams.

This infor­ma­tion often tells us more about the soft­ware plat­form col­lect­ing the infor­ma­tion and the appa­ra­tus that we’re using than the actu­al con­tent that we are con­cerned with, which is fas­ci­nat­ing to me because I like meta stuff and I like soft­ware plat­forms and all that sort of thing. I think that’s great and inter­est­ing and I would read a paper on OCR errors in Google Books and what that means. But yeah, data analy­sis often tells you more inter­est­ing things about the method­ol­o­gy than the subject.

So we end­ed up with these mas­sive epis­te­mo­log­i­cal prob­lems any­way. We weren’t avoid­ing them at all, we weren’t being tricky or sneaky and find­ing some secret way to have an objec­tive view of what we were look­ing at. We were just cre­at­ing new epis­te­mo­log­i­cal prob­lems. We were just mov­ing things around.

I often equate data analy­sis with cri­tique as a prac­tice. They’re both based on decom­pos­ing things, break­ing them out into parts, fig­ur­ing out how they work. I want to talk a lit­tle bit about this idea of com­po­si­tion over cri­tique. When I say com­po­si­tion, I mean this in the sense of Bruno Latour’s ​“Compositionist Manifesto” pub­lished back in 2010. When Latour talks about com­po­si­tion, he sets it up not in oppo­si­tion but sort of as an alter­na­tive to cri­tique. He comes from a sci­ence and tech­nol­o­gy stud­ies back­ground, so he’s an STS guy, and his point is that cri­tique only gets you so far. The way he puts it is if you have a sledgehammer—not a reg­u­lar hammer—or a wreck­ing ball, there’s pret­ty much one thing you can do with that, and it’s super use­ful for clear­ing out old debris and all that sort of thing. But what hap­pens when you’re done with that? I’ll quote him at length here. He’s talk­ing about the dif­fer­ence between cri­tique and com­po­si­tion. He says:

The dif­fer­ence is not moot, because what per­forms a cri­tique can­not also com­pose. It is real­ly a mun­dane ques­tion of hav­ing the right tools for the right job. [The sledge­ham­mer can] break down walls, destroy idols, ridicule prej­u­dices, but you can­not repair, take care, assem­ble, reassem­ble, stitch togeth­er. It is
no more pos­si­ble to com­pose with the para­pher­na­lia of cri­tique than it is to cook with a see­saw. Its lim­i­ta­tions are greater still, for the ham­mer of cri­tique can only pre­vail if, behind the slow­ly dis­man­tled wall of appear­ances, is final­ly revealed the nether­world of real­i­ty. But when there is noth­ing real to be seen behind this destroyed wall, cri­tique sud­den­ly looks like anoth­er call to nihilism. What is the use of pok­ing holes in delu­sions, if noth­ing more true is revealed beneath?
Bruno Latour, ​“An Attempt at a ​‘Compositionist Manifesto’ ”

This in turn reminds me of some­thing that I read ear­li­er this year in the dig­i­tal human­i­ties milieu from Mark Sample. He wrote a blog post called ​“Difficult Thinking about the Digital Humanities” and I’ll quote him as well here. I’m sort of tak­ing him a lit­tle bit out of con­text, but I hope he’ll for­give me. He says

Two hall­marks of dif­fi­cult think­ing are imag­in­ing the world from mul­ti­ple per­spec­tives and wrestling with con­flict­ing evi­dence about the world. Difficult think­ing faces these ambi­gu­i­ties head-on and even pre­serves them, while facile think­ing strives to elim­i­nate complexity—both the com­plex­i­ty of dif­fer­ent points of view and the com­plex­i­ty of incon­ve­nient facts.
Mark Sample, ​“Difficult Thinking about the Digital Humanities”

He’s sort of draw­ing a line between that facile think­ing and cri­tique, and he says, ​“I’m dis­sat­is­fied with that word ​“crit­i­cal” and all its variations—that’s why my for­mu­la­tion empha­sizes dif­fi­cult think­ing over facile think­ing.” I don’t think it’s a coin­ci­dence that at the same time that he wrote this essay, he was also in the mid­dle of an incred­i­bly pro­lif­ic peri­od of writ­ing about and build­ing Twitter bots. He wrote this great essay about bot arche­types where he talks about closed bots ver­sus green bots. @everyword is a great exam­ple of a closed bot. It takes a dic­tio­nary and it iter­ates through the dic­tio­nary and it tweets every word in that dic­tio­nary, and then it’s done. That’s a closed bot. 

A green bot is some­thing that might source from the greater world. Sample’s bot @whitmanfml is a great exam­ple of that. It takes quotes from What Whitman which are a closed cor­pus and then mix­es them with the #FML hash­tag, source from Twitter which is an open cor­pus, and that’s what he’s talk­ing about when he talks about green bots and how green bots have this capac­i­ty to sur­prise and all these oth­er things.

I don’t think it’s a coin­ci­dence that he was think­ing about these things at the same time. Another great exam­ple of a bot that he built is more of an activist bot, sort of the equiv­a­lent of— It’s activist in the sense that a Brechtian play is activist. He calls it an exper­i­ment in spec­u­la­tive sur­veil­lance, and it comes up with all these tweets about who NSA PRISM has caught doing what. If you fol­low this, it keeps you on your toes, essen­tial­ly, as you fol­low it through your feed. 

What I want I want to chal­lenge peo­ple in dig­i­tal human­i­ties to think about is how you can reverse the polar­i­ty, so to speak, on the tools of dig­i­tal human­i­ties. If you’re learn­ing things about the way texts are con­struct­ed or how they’re sit­u­at­ed, that kind of thing, and you’re learn­ing these things through analy­sis be it through data like ngrams or any­thing like that, by all means pub­lish the results of that sta­tis­ti­cal analy­sis or visu­al­iza­tions or what­ev­er. But I think it’s a real­ly impor­tant exer­cise to also try turn­ing that research on its head and using your find­ings to gen­er­ate things that go out into the world.

A real­ly enlight­en­ing exam­ple for me is the reg­u­lar expres­sion. For those who don’t know, a reg­u­lar expres­sion is a com­put­er sci­ence con­struct, it’s math­e­mat­i­cal con­struct, it’s a lin­guis­tic con­struct. It’s one of these great things that sort of cuts across all sorts of dis­ci­plines, and you can write math­e­mat­i­cal proofs that a reg­u­lar expres­sion is also a state machine from this oth­er field, and it’s also iso­mor­phic to this and that. But for peo­ple in dig­i­tal human­i­ties espe­cial­ly you might be famil­iar with the reg­u­lar expres­sion because it is one of the most com­mon tools for extract­ing text from oth­er text.

So if I want to find every set of comma-separated phras­es in a book, I could write a reg­u­lar expres­sion that essen­tial­ly says find a block of let­ters and then a com­ma, fol­lowed by an arbi­trary num­ber of blocks of let­ters and a com­ma, fol­lowed by maybe an ​‘and’ maybe not if it’s an Oxford Comma or what, and then ter­mi­nat­ing in a ter­mi­na­tor like a peri­od or an excla­ma­tion point or any­thing like that. (I’m a pro­gram­mer; it took me ten years to real­ly inter­nal­ize reg­u­lar expres­sions.) You feed that in, and then you feed in the book, and then it spits out on the oth­er end the match­es for all that. You can also use reg­u­lar expres­sions to find and replace things. So if I want to find all of those comma-separated phras­es and replace them with the word ​“meow,” I can do that very eas­i­ly with a reg­u­lar expression. 

What I had­n’t real­ized in ten or twelve years of using reg­u­lar expres­sions is that they also can be turned on their head. You can use a reg­u­lar expres­sion as a gen­er­a­tive tool. There are tools out there where you can put in a reg­u­lar expres­sion that might say— If you wrote a reg­u­lar expres­sion that said ​“find every instance of two vow­els adja­cent to one anoth­er” you could also feed it into a machine that says ​“giv­en the rule ​‘two vow­els adja­cent to one anoth­er’ gen­er­ate every string of char­ac­ters that val­i­dates that rule.” and then it would spit out a few hun­dred vow­el pairs for you. You could get weird­er than that. You can feed in sets of syl­la­bles and you could have it gen­er­ate whole new lan­guages and that sort of thing based on this tool that is nor­mal­ly used for analy­sis and decom­po­si­tion and tak­ing things apart. You can turn it on its head and make it into some­thing generative.

I want to show anoth­er exam­ple here. This is on my mind right now because this is some­thing that is hap­pen­ing at the moment. I orga­nize this thing every November called NaNoGenMo, and it’s like NaNoWriMo, which is National Novel Writing Month where peo­ple sign up and they pledge to write a 50,000 word nov­el over the course of the month of November. Last year I posit­ed that we could do NaNoGenMo where you’d pledge to write code that gen­er­ates a 50,000 word nov­el over the course of November. So we have a bunch of peo­ple who sign up and do this. Academics, pro­gram­mers, peo­ple who are nei­ther. We have 61 projects going right now. Some of them are already done. This one’s great; it’s called 50,000 Meows, and you just feed it a book and we can get the new all-meow ver­sion of Moby Dick. Much improved, I think. It retains the syn­tax and cap­i­tal­iza­tion and all that sort of stuff. So it still has the shape of Moby Dick.

I’m work­ing on some­thing where I’m look­ing at peo­ple who have stud­ied plot­ting for nov­els and reverse-engineering their analy­sis of the way plots in nov­els work to gen­er­ate some­thing approx­i­mat­ing a nov­el and part of what this does is it points out the holes in their mod­el as well. So it does sort of oper­ate as cri­tique as well. I pulled all this stuff from one of these ​“how to write a nov­el for a begin­ner” things, and this per­son has the ​“ten steps” of a nov­el. You start with sta­tus quo, then some­thing hap­pens, then the char­ac­ter makes a deci­sion to act. So I read through this whole thing and took it extra­or­di­nar­i­ly lit­er­al­ly, and I built a short first pass of this, which is 

If you’re inter­est­ed, there’s a resource thread where peo­ple post resources they think might be use­ful, any­thing from tools for nat­ur­al lan­guage pro­cess­ing to cor­pus­es of data that they think might be inter­est­ing. I learned about some­one who com­piled 10,000 plot ele­ments that could pos­si­bly be in a book, so I’m prob­a­bly going to dig into that pret­ty soon as well. To me as an out­sider, this is all dig­i­tal human­i­ties work as well. We’re ana­lyz­ing the form of what’s out there and then inter­nal­iz­ing it. In this case we’re writ­ing an algo­rithm based on that, and then spit­ting it back out in the world. 

This is the same thing as when I did @twoheadlines. This was me look­ing at a lazy form of joke on Twitter and going, ​“I could para­me­ter­ize this and just gen­er­ate this joke for­ev­er.” I don’t know if Mark Sample would con­sid­er this an activist bot, but I con­sid­er it activist for a very nar­row band of cul­tur­al activism where I want to kill a par­tic­u­lar kind of joke. And I’ve sort of suc­ceed­ed. On a week­ly basis, some­one tells a lazy joke like this and then some­one else rather innocu­ous­ly goes, ​“Oh that’s just like @twoheadlines!” And I’d like to think that they look at it and feel bad.

Much of what I do is just notic­ing pat­terns in how peo­ple com­mu­ni­cate in dig­i­tal media and fig­ur­ing out how to repli­cate that and send it back out in the world. When I see peo­ple doing just analy­sis, I’m always hold­ing my breath wait­ing for the next step, which is what are you going to do with that? What are you going to make with that? Analysis for its own sake is fine, I like analy­sis for its own sake, but at that point that I’m wait­ing for some­one else to take that and see what they’re going to do with it.

I think that’s it for the planned part of my talk here. I’m inter­est­ed in open­ing it up to ques­tions or com­ments or anything.

This is just a lit­tle bonus. I built this this morn­ing while I was work­ing on my talk because I acci­den­tal­ly built it. This just gen­er­ates a ran­dom ngram. So if you want to do some analy­sis, there’s a tool for that now. This was sort of inspired by this great web site Spurious Correlations. This algo­rithm has a zil­lion dif­fer­ent data sets and finds ones that cor­re­late well, and then it says, ​“Well here you go. Drawn your own conclusions.”

Audience: But you’re not actu­al­ly look­ing for sequences that are cor­re­lat­ed, they’re just…

Darius: No. They’re just ran­dom words from the dic­tio­nary, but some­times you’ll notice that ​“polar­iza­tion” and ​“folk­lore” are pret­ty well cor­re­lat­ed, and what does that tell us about the his­to­ry of the writ­ten word?

Audience: That’ll be for your next talk.

Audience 1: I’m won­der­ing if your nov­el gen­er­a­tion algo­rithm could be forked and turned into a schol­ar­ly arti­cle generator.

Darius: There are already tons of schol­ar­ly arti­cle gen­er­a­tors. That’s a well-trodden path. It’s usu­al­ly peo­ple from the hard sci­ences, who cre­ate gen­er­a­tors that gen­er­ate sort of decon­struc­tion­ist tomes and then sub­mit them. [cross-talk com­ment from audi­ence mem­ber] Random post-modern stuff, and then sub­mit them to actu­al jour­nals and see who they can get to actu­al­ly accept them for pub­li­ca­tion. And this has hap­pened. But it’s also gone the oth­er way. There are sci­en­tif­ic papers as well that have been equal­ly, sad­ly, accept­ed by jour­nals, too. You should Google it. There’s a lot of good stuff there.

Audience 2: My day job is over in the English depart­ment, so I’m going to try to ven­tril­o­quize a ques­tion from a kind of cranky senior col­league who, were such a per­son sit­ting here, he or she might point out that if you look at Literature with a cap­i­tal ​“L,” the kind of lit­er­a­ture that we expect to do real cul­tur­al work. You look at titles and get things like War and Peace, you get Crime and Punishment. These are the big ques­tions, right. And the projects that you’ve shown, one of them is about killing a cer­tain type of joke on Twitter. How would you address a kind of cranky skep­tic who sort of want­ed to know some­thing about the poten­tial for this form to address the big ques­tions, the human condition?

Darius: I’d show them [Last Words] to start with. This is occur­rences of the word ​“love” in last words of exe­cut­ed Texas death row inmates. This is a cor­pus that I got my hands on. Someone just point­ed me to it, and this is one of those 45 minute projects for me. I did­n’t even set out to build this, I just said ​“I have this weird set of data.” I built soft­ware that scraped out all the last words. I wrote a pro­gram that just put the entire text of peo­ple’s last words in front of me and just sort of flashed it by like flash cards. I noticed the word love appeared many many times. I said ​“what if I fil­ter for the word ​‘love?’ ” It came up with this, and I said, ​“I’m done. That’s it. It’s done. Ship it.” That was over lunch at work one day. And this has been used by actu­al anti-death penal­ty advo­cates. They use this in their work­shops. A Christian news­pa­per wrote an arti­cle about this and Christ and the death penal­ty, and can Christians real­ly sup­port the death penal­ty, that sort of thing.

So maybe not NaNoGenMo, yet. I don’t think we’re going to have War and Peace, but I do think there’s plen­ty of poten­tial to address larg­er issues.

Audience 2: And one ver­sion of that ques­tion is kind of this cranky hypo­thet­i­cal col­league, but anoth­er ver­sion is the con­cern that a lot of what we often see is either deriv­a­tive or par­a­sitic in the sense that it’s a kind of in-joke if you will. So the 50,000 meows is fun­ny, but because we have National Novel Writing Month, so you already have to be in on that as a kind of Internet meme or phe­nom­e­non in order to then appre­ci­ate this won­der­ful deriv­a­tive work, and that’s a kind of bar­ri­er to entry.

Darius: Right. NaNoGenMo itself is my lov­ing jab at NaNoWriMo because it is very strange to have an event where the def­i­n­i­tion of nov­el is ​“50,000 sequen­tial words.” So yes, I do agree there’s very often a bit of a bar­ri­er there. On the oth­er hand, I think you could make the argu­ment that what I’m seek­ing to do— It’s that whole pub­lic intel­lec­tu­al ver­sus uni­ver­si­ty and pub­li­ca­tions and that sort of thing type of debate. I’m inter­est­ed in deploy­ing things to pub­lic net­works like Twitter and Facebook and the Internet at large, where peo­ple are more like­ly to laugh at a ref­er­ence to NaNoWriMo than they are to laugh at a ref­er­ence to Foucault.

Audience 2: But ulti­mate­ly the same sort of under­ly­ing message…

Darius: Yeah. I think it’s a lot of the same work that’s being done. And much like with the debates over the pub­lic intel­lec­tu­al and their role, they can go on for­ev­er as well.

Audience 3: How do you see the field of soft­ware crit­i­cism or cul­tur­al crit­i­cism through soft­ware evolv­ing? What’s next?

Darius: I don’t know what’s next for the evo­lu­tion of cul­tur­al crit­i­cism or oppo­si­tion through soft­ware. I see more and more work done with— I was talk­ing ear­li­er about this today. In my head for a long time, I’ve had this divide between soft­ware that does work like util­i­ty soft­ware, and soft­ware that does art. You have util­i­ty bots that you can tweet at a bot and say ​“Remind me at 12pm to have lunch.” and then it would tweet at you at 12pm and say, ​“Have lunch.” That’s a Twitter bot. It’s not real­ly an art bot, I don’t think, but maybe you could make a case for it. As opposed to my Metaphor-a-minute, which just cranks out metaphors all the time, which is much more sit­u­at­ed as an art thing. 

But it’s strange because there are also activist bots. I’m host­ing some­thing called Bot Summit on Saturday [November 8, 2014] in Boston, where a bunch of prac­ti­tion­ers are get­ting togeth­er both online and in per­son to talk about this sort of stuff. One of the things we’re going to talk about is this whole ques­tion of activist bots. There’s a strange ten­sion between activism through bots and the terms of ser­vice of places like Twitter, which don’t let you unso­licit­ed reply to peo­ple. So I’ve had peo­ple say, ​“Can you help me make a bot that finds peo­ple tweet­ing bad facts about cli­mate change and gives them the data to help them come around on it.” And I tell them no, I can’t do that because that bot will get banned auto­mat­i­cal­ly by Twitter as spam, because tech­ni­cal­ly it is spam. They did not sign up to get harassed by an enti­ty for things that they are tweet­ing. So it makes it very hard to do activist work on Twitter, but with­in activism you have Mark Sample’s NSA PRISM bot, which is activism. It is also art. I drew a con­nec­tion to Brecht on pur­pose there. It’s pos­si­ble to do art and activism at the same time. It’s also pos­si­ble to write some­thing on a sign that says ​“Down with So-and-so” and go out in the street, and that is activism as well, and that’s not art.

So I’m very inter­est­ed in the future of activist bots and activist soft­ware. There’s some­thing that some­one built called Block Together, where you can as a com­mu­ni­ty come up with a black­list of peo­ple who are tox­ic to your com­mu­ni­ty. And then every­one in your com­mu­ni­ty can authen­ti­cate with Block Together and press a but­ton and simul­ta­ne­ous­ly block all 5000 of those peo­ple. So mem­bers of your com­mu­ni­ty can be pro­tect­ed from hav­ing to deal with mem­bers of anoth­er com­mu­ni­ty they might find tox­ic. But it stays with­in the bounds of Twitter’s terms of ser­vice. That’s very inter­est­ing work as well.

In terms of trends, I’ve noticed a lot more visu­al things hap­pen­ing on Twitter recent­ly. There’s ⋆✵tiny star fields✵⋆ by Everest Pipkin, which has 20,000 fol­low­ers, blows away any of my cre­ations, and just tweets these beau­ti­ful, serene lit­tle visu­al com­po­si­tions. And I love @ARealRiver, which tweets indi­vid­ual things that look like a sort of scene by a riv­er. As you scroll down you’ll notice it’s con­tin­u­ous, and you can just scroll for­ev­er and fol­low this riv­er. There’s a boat in there, there’s a horse. If you go all the way down, you’ll hit a moun­tain range that per­sists for quite some time, and then there’s met­ro­pol­i­tan areas it’s passed through. It’s a per­sis­tent land­scape on Twitter. It’s very inter­est­ing and it’s very excit­ing to me. So I’m notic­ing a lot more visu­al stuff play­ing with emo­ji as well, but I can’t pre­dict the future. Hopefully I get a bet­ter sense when we have Bot Summit this weekend.

Audience 5: I want­ed to ask you about the term ​“Weird Internet.” Can you give us the back­ground on that term, and also talk more broad­ly beyond Twitter where you see the weird Internet?

Darius: Sometimes when I say Weird Internet, some peo­ple thing I’m talk­ing about Weird Twitter, which is its own sub­set of sur­re­al Dadaist humor that hap­pens on Twitter. But when I say Weird Internet I mean that very specif­i­cal­ly. I’m talk­ing about almost a return to a folk Internet where there’s a place for things that aren’t just san­i­tized of your birth­day par­ty that your grand­moth­er can click Like on. People have recent­ly referred to it as like the return of Web 1.0. There was an arti­cle recent­ly about that. I don’t know if you’ve seen tilde​.club but you sign up—there’s a huge wait­list now, like eight thou­sand peo­ple or some­thing on the waitlist—but I think Paul Ford set this up and it’s lit­er­al­ly just… I don’t know if uni­ver­si­ties do this any­more, but I know they did when I was in school where it’s like, ​“Okay here’s your shell account, and here’s the Unix serv­er and your user­name is blahblahblah.edu/~dkazemi” and then you can upload what­ev­er you want on there and host it on there. This is an attempt to recre­ate that, and if you go to peo­ple’s sites […] it’s got this lo-fi look to it. But beyond that, it’s got a lo-fi appa­ra­tus as well. One of my friends was say­ing ​“I have friends who are com­plete­ly not tech­ni­cal ask­ing me for help with ssh because they have a tilde​.club account and they want to update their site.” I con­sid­er this Weird Internet, I con­sid­er bots, espe­cial­ly non-utility bots to be Weird Internet. Anything that makes you stop in your tracks or sur­pris­es you at all I would con­sid­er part of that milieu. I just want strange and unex­pect­ed beau­ti­ful things to hap­pen all the time.

Audience 6: I don’t know this for a fact, but it was my impres­sion that that start­ed out as a joke kin­da sim­i­lar to how some of the things you talked about have a com­i­cal aspect to them, but Paul was quick­ly over­whelmed by all these peo­ple that want­ed to use that site. There were some peo­ple that I think were part of the joke, like Matt was talk­ing about being on the inside—

Darius: Right. Some of it’s almost like an iron­ic, hip­ster, ​“I’m gonna use ssh in addi­tion to build­ing a bike out of my woodshed.”

Audience 6: But then there were all these oth­er peo­ple that got in on it because they actu­al­ly liked the idea of it. Like, ​“Actually, no, I don’t mind ssh’ing to this com­put­er and cre­at­ing an HTML file and hav­ing a web site there.” And peo­ple that are redis­cov­er­ing that that was the way that the web start­ed. There’s a new gen­er­a­tion of peo­ple now that are [observ­ing?] this is actu­al­ly how it began.

Darius: Myself, over the last few years I’ve been slow­ly extri­cat­ing myself from blog­ging plat­forms like WordPress. I archived my WordPress site and it exists in sta­t­ic form now but I can’t log in, there’s no data­base or any­thing like that. Instead if I want to cre­ate a blog post, I open up a text edi­tor and I type “” and then I start like I used to when I was in 7^th grade and mak­ing lit­tle web pages. So now my essays look more like this. I just write it all by hand in a text editor.

But it also gives me the free­dom to do things like embed soft­ware and make it way more inter­ac­tive than it oth­er­wise would’ve been because on WordPress it’s like, ​“Well can I find a plu­g­in that does the thing that I want? Do I have to write a plu­g­in?” As a tech­ni­cal per­son, this gives me more free­dom. But I think also even if you’re not that tech­ni­cal and you’re just learn­ing ssh, it can be exhil­i­rat­ing to have— I remem­ber when I was a teenag­er I think, and I first learned that I could log into a com­put­er remote­ly some­where else, log into my home com­put­er from some­one else’s com­put­er and get files off of it, that was this incred­i­ble, exhil­i­rat­ing thing for me not just because of that con­nec­tion, but also it was this thing that I con­trolled. I could lit­er­al­ly install soft­ware on my home com­put­er from some­one else’s machine, and that was very cool and it felt empow­er­ing at the time, and I think you get a whiff of that by join­ing some­thing like tilde​.club as well, even though it’s more man­aged. But it’s cer­tain­ly less man­aged than Facebook, for example.

Audience 6: I guess what I meant to ask was have you noticed this sort of arc in some of the work you’ve done where it start­ed out with this ini­tial pur­pose of doing some­thing maybe with a com­ic val­ue or what­ev­er, but then like with tilde​.club it took on this oth­er mean­ing. Like you start­ed out with a par­tic­u­lar idea but it sort of mor­phed into some­thing else that had more use than you anticipated.

Darius: Definitely. A great exam­ple of that is Museum Bot. I came home from work one day and my wife was already home, and she said, ​“Hey, I don’t know if you saw, the Metropolitan Museum of Art released 400,000 high-resolution images along with meta­da­ta about its deep, deep back cat­a­logue.” And I as like, ​“That’s real­ly cool!” And she was like, ​“You might want to do some­thing with that.” And I was like, ​“Yeah, that would be cool to do.” And I sat there for a half an hour, forty-five min­utes, look­ing at their col­lec­tion and fig­ur­ing out what I could do with it, and I could­n’t come up with any­thing good. I was just like, ​“I can’t think of any­thing inter­est­ing. But you got me all excit­ed so I guess I’ll just build the most bor­ing pos­si­ble degen­er­ate case.” Which is it just grabs some­thing at ran­dom and tweets it. 

At the time I was mak­ing it, it was a throw­away. But because it’s this autonomous agent that goes out into the world and is sort of con­sis­tent­ly sur­pris­ing. The Met’s col­lec­tion often has, I mean you know, ​“Petals of a Composite Cornflower Pendant,” that’s weird; An unla­beled just…vase; a drink tum­bler, like for your whisky. I would not have expect­ed to see this stuff in the cat­a­log at the Met. People real­ly liked this, and it also says things about cura­tion. People can look at this and look at what’s on dis­play at the Met and go, ​“Oh. What’s on dis­play at the Met is noth­ing like what a true ran­dom sam­pling of items in the Met’s cat­a­log is.” So peo­ple start to learn things about cura­tion by engag­ing with this bot, and then for some peo­ple it’s almost a sta­tus thing. It’s like, ​“Oh, I’m sign­ing up for cul­ture.” It’s like lis­ten­ing to NPR so you can tell your friends that you lis­ten to NPR. ​“Have you heard the lat­est Serial episode? Come on.”

So there’s a lot going on here, and I had zero zero zero idea when I built it. I built it and I had din­ner, and I was like okay, I can for­get about this for­ev­er. And it keeps com­ing back up with more and more peo­ple who are inter­est­ed in it and want to talk about it and so fort. 

Audience 8: I want to go back to your exam­ple where you were say­ing you sort of drift­ed away from exist­ing plat­forms and start­ed build­ing your own stuff includ­ing blog posts and tied that to the activism stuff that you were talk­ing about ear­li­er. Of course, activists have real­ly embraced the Internet includ­ing some of the weird stuff as well. But they will obvi­ous­ly and prob­a­bly do see the val­ue in mov­ing off plat­forms like Facebook and Twitter because those are owned by pri­vate com­pa­nies and they exert their con­trol over algo­rithms, over stuff they cen­sor over their terms of ser­vice, and activists don’t like that because that gets in their way. On the oth­er hand, as soon as they want to move away they face a dif­fer­ent bar­ri­er, which is they need to learn how to make this stuff, and a lot of them—the major­i­ty of them, probably—don’t know how to do this stuff. So what would you say is like…there’s obvi­ous­ly kind of a good thing about more peo­ple learn­ing about the Weird Internet, but there’s also this bar­ri­er of hav­ing skills to build their own stuff. And also then let­ting oth­er peo­ple know that stuff’s out there, because unless its on Facebook very few peo­ple will learn that it’s there because it’s just some lit­tle web site.

Darius: To the sec­ond point first, I think that’s unavoid­able. I think you see that ana­log in just tra­di­tion­al activism as well. As an activist, you both need to be able to go out into a pub­lic square and protest and maybe do it by get­ting a per­mit and fol­low­ing local laws, or maybe not so much. But you also need a pri­vate space in your own home or a build­ing some­where or a pub­lic park or some­thing, where you can just gath­er and do some­thing a lit­tle more pri­vate on your own terms as well. So to me, I see that as a loose ana­log to run­ning your own serv­er and ser­vices ver­sus going out into the world. I don’t ever want to stop putting my stuff on Twitter as a plat­form because whether we like it or not, that’s sort of like the pub­lic forum, or one of them.

On the edu­ca­tion gap, I agree with you com­plete­ly. I think I’m eth­i­cal­ly oblig­at­ed, or moral­ly oblig­at­ed to teach peo­ple how to do this kind of stuff, so I pub­lish the source code for almost every­thing I do. But pub­lish­ing source code is like what­ev­er, it does­n’t help 99.9% of peo­ple. So I write prose tuto­ri­als, I run work­shops, I run things like Bot Summit, I go speak at places, that sort of thing. That’s a per­son­al choice that I make as a cre­ator, to actu­al­ly give peo­ple these tools that’ve helped me so much in me doing those things, because I like see­ing oth­er peo­ple’s work and what they do with it.

Audience 9: Are you not afraid that at some point there will be so many peo­ple doing Weird Internet that it’ll stop being weird?

Darius: For a long time I was the only per­son doing game ana­lyt­ics and then every­one was doing it, and I moved onto Weird Internet. And I sup­pose if I real­ly feel crowd­ed out I’ll just move on to some­thing else.

Audience 10: I noticed that when you talk about your bots, often you refer it’s like, ​“Oh it took fif­teen min­utes, it took me forty-five min­utes.” I was won­der­ing how you wish your audi­ence get out from that kind of comment.

Darius: I think I men­tion that because I con­sid­er mak­ing a Twitter bot on the same lev­el as most oth­er cre­ative endeav­ors, so I’m a big fan of— There’s a guy who I met recent­ly, Jonathan Mann, he does a song a day on YouTube and has been doing it for many years. He passed the 2000 song mark recent­ly, and much like me not even close to a major­i­ty of his stuff is a big hit or any­thing, but [we] prob­a­bly have rough­ly about the same hit rate, about one out of every fif­teen things that we make gets some kind of response oth­er than the usu­al back­ground noise of like, ​“Oh that’s cool.” 

I have this blog post I wrote called ​“Thoughts on Small Projects” that goes in detail on this. I think it’s very impor­tant as a cre­ative per­son to get to a place where not only have you skills got­ten to a cer­tain lev­el where you can cre­ate things very quick­ly, but also your expec­ta­tions have been brought down as well. I think it’s very impor­tant to do. I spent until I was maybe 27 years old or some­thing ago­niz­ing all the time over whether what I was doing was good enough, and I just sat on so many projects and had so many half-finished things sit­ting in fold­ers. I don’t have half-finished things that sit in fold­ers any­more because I decide that they’re fin­ished and I just put them out there. It’s been cre­ative­ly reward­ing for me, and I hope that oth­er peo­ple can take stuff away from that, too. That’s the great irony or big joke, that Museum Bot, which I made in an hour or two is more pop­u­lar and I get more peo­ple say­ing they liked it and all that sort of stuff than the book that I spent a year writ­ing. I’m still proud of that book, but it just did­n’t hap­pen to be what was right at the time or what­ev­er. I don’t know. 

Mostly when I talk about how quick it is, I just want to encour­age peo­ple pas­sive­ly to do what­ev­er it is they do quick­ly, whether it’s mak­ing comics or what­ev­er. I don’t mean to show­boat ​“I’m so fast and this is so easy,” but also I try to pro­vide the tools to make it easy for peo­ple as well. The rea­son I can do Twitter bots quick­ly is I have a tem­plate, which is a pro­gram that I run and it just asks me a set of ques­tions about what soft­ware I’m going to need and then I press enter and it gen­er­ates a place­hold­er Twitter bot that just tweets ​“Hi” every hour. Then I don’t have to write any of the scaf­fold­ing. It’s like here’s all the scaf­fold­ing, now you can just work on the core of it. That saves me sig­nif­i­cant time on every project as well. But I made sure to pub­lish this and I know oth­er bot mak­ers who use this, and it saves them time as well. So it goes back to that moral oblig­a­tion that I feel to enable oth­er peo­ple to be able to do this stuff.