Weavrs - autonomous Twitter bots | Hacker News Reader

Weavrs - autonomous Twitter bots

weavrs.com

43 pointsjaxonrice40 comments

That site got a fair bit of press (and not in a good way) after they set up one of those bots pretending to be the author Jon Ronson. Jon interviewed the people behind it, judge them for yourself: http://www.guardian.co.uk/commentisfree/video/2012/mar/27/jo...

that was one of the funniest things I've ever seen, they certainly take an interesting position

The Weavr guys have posted the backstory of the jon_ronson bot on their blog. It's an interesting read.

http://philterphactory.com/2012/04/02/jon_ronson-a-weavr-of-...

Not sure if it's just me, but those guys come across as extremely unlikeable in that video. I don't get how they seem oblivious to why Jon is upset about it, and even at one point one gets (faux?) annoyed that real Jon is telling people following fake Jon to stop doing so.

The central guy comes across as a total psychopath. Twisting everything he says in a strange way. Almost saying the fact that he uses his name on Twitter makes him self-obsesed. I've just looked at my list of @replies and the last ten are all users using their real names.

They have some of the most absolutely weakest arguments I've ever heard yet get away with it. It's as if they don't care if the thousands watching the video aren't convinced at all but if they can annoy Jon they have succeeded.

He says the reason he annoys people by making these accounts is because "I want to understand the algorithms that are used by wall street". These aren't really 'robots', they're just scraping wikipedia and pushing out terms in tweets it believes are relevant, which this has shown aren't.

"I don't know how you relate to the economic collapse". I don't understand this guy, I don't know if they're acting but if they're serious then I'm worried.

The video has almost some Troll bait feeling. Found it highly amusing how concerned that guy becomes about the communication of his non existent eating habits on Twitter.

They were just making fun of Jon Ronson. I think their arguments hold up just enough to make Jon's request to kill the bot sound a little absurd. That seems to be their only objective in the video. I thought the bit about trading algorithms and how it relates to what they're doing was a little over the top. The analogy to photography made by the guy on the left at the end of the video was genius.

http://en.wikipedia.org/wiki/Reductio_ad_absurdum

I take it they force themselves to come across as extremely unlikeable but do throw in that they want to show how an inverted stock put compensation Algo is not good for anyone, or at least that is how they whitewash it. While picking on a very soft target. Not being able to express themselves coherently doesn't help either.

Just make a half assed bot and ride the wave in the most nonsensical level you are able to perform at. Tip, it's not even interesting at all.

Someone could of have set up that video like this : "Three twats playing clueless to identity theft, manifest in sync when the interviewer calls them trolls".

That was priceless and very revealing of how they want to be regarded. Yet fail so hard at it. Guess neuromarketing was too crowded for these three not so luminous lightbulbs.

At times it strikes like they really believe they are on to something that is worth being "discovered", kind of hard when pushing glorified irc bots around but whatever. The twat in the middle was the funniest, guess they went for a soft target to spare him from contempt of court charges in a identity theft case... you know, for arguing of sitting arrangements. This stuff is priceless, niche priceless.

I find these guys extremely likable. Its obvious that they're aware Jon is upset and they're toying with him. They are challenging his desire to use flimsy social media to represent himself. They also challenge his right own his online identity. I feel that their attitudes are warranted.

I don't feel as though these three are merry pranksters just out for a laugh. It's obvious that something about their work kind of rubs us the wrong way. It devalues our online representations, and that's precisely what they're after. They don't glorify their algorithms. Doesn't it sting a little that their simple web bots can emulate a human fairly accurately? Doesn't it really suck when this online persona steps on our toes? Is there no way to tell the public who the real me is without makings lots of videos, online profiles, status updates and tweets.

It's all a little trivial and those three make it obvious.

I dunno, I felt like they dodged questions and twisted Jon's words which makes me think of slimy politician-speak. This coupled with their attempts to spin it as some kind of academic pursuit and the paranoid "you're trying to control us" intro just made me uneasy.

I don't really have a strong opinion on the actual bots themselves, hell I wrote one to ape the physicist Brian Cox (http://blog.mclemon.cz/notbriancox-i-do-some-stupid-things-w...) after reading about Markov Chains one lunchtime - though I only posted a couple tweets and made it very clear it wasn't impersonating him.

It's like the guy in the center is so wrapped up in programming bots that he's taken on robot-like qualities himself (lack of empathy, can't understand emotions). Positions or arguments that would make sense to me, to him are "weird" or "psychologically interesting".

The counter-argument Mr. Ronson could have presented would be: "well if someone goes on twitter, and intends to follow me and they end up following your jon_ronson bot instead, that is fundamentally dishonest to both me (the real Jon Ronson) and the follower".

Unlikeable inasmuch as the nihilists from <i>The Big Lebowski</i> would be unlikeable should you meet them in reality.

This video is at least very thought-provoking.

Oh damn guys, this is epic. I see what you're doing here. User-generated bots with unique personalities. Generate enough and put them into the wild that is twitter, and you'll quickly find what algorithms are best at "befriending" most people.

I'd always predicted pure algorithmic AI would emerge only after we'd stitched together the human hiveAI (a small 7 billion neruons), but it may just grow with it.

Here's one (of many) thoughts: imagine generating a twitter bot (sorry, weavr) directly from your search and/or browsing history...

Everything about this seems stupid. In the video bazzargh posted these guys come off as absolute pricks, oozing with smugness and condescension and completely devoid of empathy. They use (steal) someone's identity (name + pic), the person asks them nicely to stop and they insult and ridicule him: "You'd like to kill these algorithms; you feel threatened in some way." These guys have such a laughably asinine narrative, I can't believe they deliver it with an (almost) straight face. Why didn't the creators use their own identities? Why didn't they stop using Ronson's when it became clear it was irritating him?

To the software output: Maybe I'll eat my words but it seems to be complete garbage. Basically boilerplate, randomly generated spam. I typed "wine" into the "find weavers" box and picked the top result: a weaver/spambot called "In Vino Veritas" (in wine, truth) with a wine glass as an avatar. This should be a good match for wine, I figured! I was asked if I'd like to "chat" with the bot and did so, I pitched it a softball: "What is your favorite wine?" It came back with "I don't know much about wines but I prefer those from California." "In Vino Veritas" doesn't know much about wine? Not exactly crushing the Turing test here... http://screencast.com/t/fKGzERNr

I looked at the bot's profile page and it had a post "I'm dreaming something about #proof and #evidence." In the interview, Ronson says that the bot posted 'Dreaming about #time and #cock,' almost the exact same string with the hashtags substituted.

    printf("Dreaming of %s and %s", x, y);

And we're supposed to regard this as what, cutting edge AI? Winebot that can't say anything about wine, bots feeding random values into preformed strings and tweeting them, taking a word from some interaction (e.g. "wine"), pulling a pic from google and posting to tumblr... nothing here is remotely interesting or new. If the creators weren't such class-a assholes I'd be more polite in my assessment, but as they can't be bothered to care about how their actions affect others, I don't feel the need to. They've done their controversy-color-by-numbers thing and got their publicity, the product (?) is worthless, the sooner their 15 minutes passes the better.

I don't think anyone is bragging about the algorithms here. No one claims it passes the Turing test. The AI isn't the real point. Furthermore, since when does someone's name + picture pass as their identitym That's exactly what this trio is pointing out. After doing some research, it appears that these guys in the video are not programmers, but artists and professors. http://zeroinfluence.wordpress.com/projects-archive/ http://tedxtalks.ted.com/video/TEDxManchester-Dan-OHara-Skeu... http://www.danohara.co.uk/cv.html

If we continue this critique, can we stop judging the quality of their algorithms (which they're probably not interested in) and instead critique the artistic value of emulating human online behavior?

"can we stop judging the quality of their algorithms (which they're probably not interested in) and instead critique the artistic value of emulating human online behavior?"

I actually addressed this first, before the software output. They are being jerks and not demonstrating anything interesting. Please tell me: what is the artistic value?

'Thru this work we now know it's possible to piss people off by putting their name and picture on a spambot.' Is this it? Don't get me wrong, I think creative, high quality trolling can be considered art in its highest form (see Andy Kauffman, Sacha Baron Cohen, 4chan at its best, etc.) but simply slapping someone's name on a spambot is uninteresting, unoriginal, childish, and above all, uncreative. It provokes no thoughts beyond "wow, these guys sure are jerks."

For what it's worth the creators mention "the algorithm" constantly in their interview and seem to ascribe it some value, so I think it's a fair subject of discussion & critique.

I think creative, high quality trolling can be considered art in its highest form (see Andy Kauffman, Sacha Baron Cohen, 4chan at its best, etc.) but simply slapping someone's name on a spambot is uninteresting, unoriginal, childish, and above all, uncreative

I disagree with that idea. In my view the level of trolling is defined not by the troll, but by those being trolled. Obviously the guy in the video is taking their relatively unsophisticated level of trolling as a serious matter. Just by doing so their trolling is in my view justifiable, as it achieves its objective. Also I personally find the way they were approaching this interview quite amusing, i.e. they are shifting in a concern-trolling mode, almost taking a therapeutic stand. This lets the interviewer argue to a certain degree against a wall as they try to imply some deeper motivations by the interviewer, and find the "hidden" intentions for why he becomes so concerned about a bot spouting random information - that apparently still enough other people take serious enough to communicate with.

I'm still a little caught up in the "identity theft" thing. Names and appearances are not the limits of one's identity. If this were the case, halloween would be the criminal's holiday. The trio, perhaps inadvertently, point out a fundamental flaw with both twitter and how the user approaches it. If someone cannot differentiate themselves from a bot spitting out ("dreaming of %x and %y", x, y) then is there really an identity to steal? This man is fighting for his unique right to present 140 characters, a string of characters (his name) next to 5329 pixels captured at one specific moment in his life, all grouped together? By what virtue does he own those rights?

Don't get caught up to the point you write about it with a damage control approach.

"If someone cannot differentiate themselves from a bot spitting out ("dreaming of %x and %y", x, y) then is there really an identity to steal?"

Is this turning to some introduction to awkward online eugenics or what... You don't know that impersonation isn't something a creative commons license on a photo from wikipedia was made for ? Do they have the right to create a substitutive persona to establish contacts with others ? Was there a change to the law and no one sent the memo or something... Differentiation is intrinsic to identity, you don't have to display it like a onus. What are you on about ? So a bot with your name on it, that would tweet "comments" on the weather and caviar, would render you non existant, voiding of the hability to contest impersonation ? That just has that touch of the same non sensical dribble that funnyjunk lawyer has, where are you going with that.

Natural rights cease because someone made a bot ? Whisky Tango Foxtrot. Hold the internet presses people.

"By what virtue does he own those rights?"

Maybe because when he signed up for the service he had those rights then and didn't give them away. Yeah, probably because of that. People accept terms of service when signing up to online services by their own predisposition and free will. They enter a contract, see where this is going. Bots that impersonate people that accept terms of service don't have precedence just because some pseudo experimental trio thought it would be a good project. They don't own his rights. That is the framing that causes the problem and why they are at fault when trying to turn the discussion to some indefensible position. They failed miserably at it.

That halloween criminal's holiday day... honestly, if you had to go make an account just to sprinkle this conversation with that piece of golden web dust/insight, you should have saved yourself the trouble. Honestly there are other places for that sort of thing. In the end it is really funny how people botching a project make this sort of thing pop up. Had a laught, but like a little britain comedy series used to state: "Computer says no."

How is this different from somebody discovering for the first time Wikipedia and becoming all defensive, that there is an article on him, that they are using a picture of him and that he didn't approve that one personally. As far as I can see, they put in the bot-about a clear line that it was a Wiki based bot.

Especially since the comments were so nonsensical one could even argue that the bot was intended to persiflage the nonsensical Tweets by pseudo-celebrities. To bring the argument to its final destination: People responding to the fake bot, are similar to those, who were watching Chaplin's The Great Dictator and wondering why Hitler movies were showing in cinemas across the USA.

@novalis, your points are a little unclear. I am unclear what "Differentiation is intrinsic to identity" means. A bot with my name and image would not render me non-existent, but it might seriously compromise my online identity. I view that as a problem. I check a radio box to some terms of service when creating an online profile, and so must a bot. But because I _understand_ the words does that mean I am more protected than the bot? They don't own his rights, but does the human? The terms of service are just words, which describe laws and agreements humans made. These are not unbreakable, infallible laws. Furthermore, the bot is certainly not subject to these laws, perhaps the creators are. But then again, is a mother responsible for the crimes of its child. I realize this metaphor does not hold up ( the trio designed the bot to be nefarious by nature ).

They did not fail miserably. They seriously perturbed the man whose identity the bot assumed because his identity weak in the first place. The mere fact that the man was perturbed I think means a victory for the trio. The online (crappy) bot gives the man a run for his a run for his money, and that's disconcerting. If the bot was so ineffective, or the project so pointless, why then is it such a problem that the bot impersonates another human. What's _really_ the problem here? What are we all upset about? Once again, If I pretend to be Steve Jobs, where is the problem. If I pretend to be no name author, where is the problem? If a bot dresses up and pretends to be a no name author, what infraction is it making?

They keep reframing it to cover the vacuum, going to the point of sugesting the photo license from the wiki page entitles them to do what they want and that the portrayed person doesn't understand it, this borders lunacy if it wasn't so obvious how poorly prepared they are. Like the software one day appeared out of nothing, had they been able. No pseudo eclectic goal post moving will explain this mess in the end and the consumation of all this adds up to a waste of time. Hope they pull the most important lessons out of this and reframe their efforts to fields they are able to work in.

I didn't like that the weavr bot posted my physical address. I thought giving access to my location would allow it to work in my city, not my actual address. And there is no way to delete that first post.

Did you set it as the Weavr's home address? Change it, and the original address will soon be out of sight and out of mind as the Weavr posts more.

No, I went to the page to create a weaver and it asked me for my physical location. I wasn't informed that it would be used in that way and I am given no way of deleting that information. Being out of sight and out of mind isn't really good enough.

Please contact support and they'll see what they can do -

https://getsatisfaction.com/philterphactory

I'd like to think these things would work because the AI is so smart that they make an intelligent contribution to twitter. Sadly I suspect they might work because it won't take much to improve upon the average tweet.

Your homepage needs some work, I'm still unclear on what Weavrs is.

Agreed - but we're still unsure what they are too. That's why we develop with them.

Double stealth or triple stealth?

"The public doesn't know what we do and we don't either ... and neither does our product."

"Triple blind testing" as KK would say. http://longnow.org/seminars/02006/mar/10/long-term-trends-in...

Triple-blind experiments will emerge through massive non-invasive statistical data collection— no one, not the subjects or the experimenters, will realize an experiment was going on until later. (In the Q&A, one questioner predicted the coming of the zero-author paper, generated wholly by computers.)

Totally agreed. It's a bunch of words about how awesome Weavrs are and calls to action to make them but I still don't know what they actually are. Show some pictures or something.

Hm, I just realized that I don't like to log in somewhere with my Google account. Too much vital stuff depends on it, and if a site forwards me to a Google login prompt there is always the danger of phishing. Couldn't you provide a classic sign up form? Granted, I may be the last person on earth to use those...

Also, why not sign in with Twitter (though I'd really prefer an independent sign up form)?

Hello. We didn't want to put people through a sign-up process AND then then ask them to fill out Weavrs profiles, so we opted for a 3rd party authentication. If we had used Twitter for Auth, then it becomes very confusing. There's Twitter accounts for Weavrs and for the User. If we had used Facebook, then people get really worried about privacy and Weavrs are not private beings. Nothing gets pulled from your Google account - just your username which you can change in Weavrs.

Getting a generic 500 Server Error when I try to sign up or sign in, FYI.

Why would using Twitter for authentication be confusing? I think it'd be great.

At first I wasn't sure what to use this for, but as soon as I saw a bot's homepage (http://lojze.weavrs.info/#/view/grid/) it dawned on me that this is the best way to get cool recommendations for stuff online.

At least I think you could be able to tune a bot so it suggests exactly the things you'd like to read.

You can also add new functionality from the Prosthetic Store. http://prosthetics.weavrs.com/

or roll your own http://developer.weavrs.com/

Reply on news.ycombinator.com