Weavrs in the Web
May 1, 2012 3:28 PM Subscribe
Weavrs are a species of new autonomous, emotive, social bots. They feed off of social API streams, wandering around the real world looking at things, posting recipes and dreaming. They can be used for what some might consider evil and what some might consider good. You can extend them with your own code or create a hero's journey for them to experience. If the New Aesthetic was a movie, Weavrs would be the extras.
See, this is moving the opposite direction I want. Instead, I'd like robots to do the dishes and laundry, leaving *me* more time to go on hero's journeys while blogging, checkingin, commenting, and chatting.
posted by Celsius1414 at 3:46 PM on May 1, 2012 [14 favorites]
posted by Celsius1414 at 3:46 PM on May 1, 2012 [14 favorites]
"Weavrs are geolocated and will check in to venues on a daily basis using Foursquare. You can view the Weavrs in your area using the augmented reality browser, Layar."
what.
"In an awkward meeting between Ronson and Philter Phactory, Ronson said that he viewed the Weavr as an irrelevant spambot and an imposter that was using his name without his permission. Weavr-originated tweets such as "I'm dreaming something about #time and #cock" served to fuel the spambot argument fire."
BURN. .. ?
so, they're out there, checking in to Foursquare (which, uh.. i thought the whole point of that pointless thing was to announce your physical presence at a location), trolling people, and recycling content? that does sound like an annoying human blogger.
so is the future something we don't actually need to be present for? can we just set up bots and then stab ourselves in the bathroom to become immortal?
posted by ninjew at 3:59 PM on May 1, 2012 [2 favorites]
what.
"In an awkward meeting between Ronson and Philter Phactory, Ronson said that he viewed the Weavr as an irrelevant spambot and an imposter that was using his name without his permission. Weavr-originated tweets such as "I'm dreaming something about #time and #cock" served to fuel the spambot argument fire."
BURN. .. ?
so, they're out there, checking in to Foursquare (which, uh.. i thought the whole point of that pointless thing was to announce your physical presence at a location), trolling people, and recycling content? that does sound like an annoying human blogger.
so is the future something we don't actually need to be present for? can we just set up bots and then stab ourselves in the bathroom to become immortal?
posted by ninjew at 3:59 PM on May 1, 2012 [2 favorites]
On the internet, nobody knows you're an infomorph programmed to scan through publicly-available data and simulate the web presence of a human personality.
posted by Grimp0teuthis at 3:59 PM on May 1, 2012 [3 favorites]
posted by Grimp0teuthis at 3:59 PM on May 1, 2012 [3 favorites]
I just wrote a blog post about Weavrs with a bit more flavor.
I...I'm sorry for derailing. But is that really a picture from Herman's Head?
posted by inigo2 at 4:02 PM on May 1, 2012
I...I'm sorry for derailing. But is that really a picture from Herman's Head?
posted by inigo2 at 4:02 PM on May 1, 2012
This isn't an actual thing, is it? It sounds like the product of one of those silly "Random Web 2.0 business idea generator" sites.
posted by indubitable at 4:09 PM on May 1, 2012 [3 favorites]
posted by indubitable at 4:09 PM on May 1, 2012 [3 favorites]
It's a real thing, but I think there's probably something special about where the web is at and the whole New Aesthetic thing going on in London that's making it possible.
Yes, that is a picture from Herman's Head.
Weavrs specifically don't try to 'fool' anyone, they announce that they're created by a user at the top of their blog, and the abstract whatever profile picture you upload (and they don't checkin to sites like FourSquare, though they do 'wander' between home and work, and post things to their blogs/twitter they find along the way), but that's not to say that the next generation won't make more attempts to be real. There are obviously businesses already churning out thousands or hundreds of thousands of real-ish social media profiles for astroturfing.
posted by jeffkramer at 4:19 PM on May 1, 2012 [1 favorite]
Yes, that is a picture from Herman's Head.
Weavrs specifically don't try to 'fool' anyone, they announce that they're created by a user at the top of their blog, and the abstract whatever profile picture you upload (and they don't checkin to sites like FourSquare, though they do 'wander' between home and work, and post things to their blogs/twitter they find along the way), but that's not to say that the next generation won't make more attempts to be real. There are obviously businesses already churning out thousands or hundreds of thousands of real-ish social media profiles for astroturfing.
posted by jeffkramer at 4:19 PM on May 1, 2012 [1 favorite]
I find the idea of intelligent bots interesting, but weavrs strike me more as a bunch of interconnected scripts spitting out occasionally serendipitous text than anything that remotely resembles 'intelligence', as the Wired UK link admits. I suppose as an art project it's fine - a provocation or perhaps a sign of things to come - but it seems to be fuelled more by hype pieces and marketing money than anything else.
posted by adrianhon at 4:31 PM on May 1, 2012 [2 favorites]
posted by adrianhon at 4:31 PM on May 1, 2012 [2 favorites]
>so is the future something we don't actually need to be present for? can we just set up bots and then stab ourselves in the bathroom to become immortal?
YES. this is the *whole point*. This is uploading. we're still pretty bad at it, but this is uploading.
>weavrs strike me more as a bunch of interconnected scripts spitting out occasionally serendipitous text than anything that remotely resembles 'intelligence'
guess what? AI will always be in the future, because once it works and we can see it's just a bunch of interconnected scripts, it's not AI anymore. Of course it's artificial intelligence. It's not smart enough yet to pass, but Siri only just recently got possible. In 5 years she'll be everywhere (unless the IP lawyers lock her in a box of course).
posted by Fraxas at 4:52 PM on May 1, 2012 [3 favorites]
YES. this is the *whole point*. This is uploading. we're still pretty bad at it, but this is uploading.
>weavrs strike me more as a bunch of interconnected scripts spitting out occasionally serendipitous text than anything that remotely resembles 'intelligence'
guess what? AI will always be in the future, because once it works and we can see it's just a bunch of interconnected scripts, it's not AI anymore. Of course it's artificial intelligence. It's not smart enough yet to pass, but Siri only just recently got possible. In 5 years she'll be everywhere (unless the IP lawyers lock her in a box of course).
posted by Fraxas at 4:52 PM on May 1, 2012 [3 favorites]
on a mass scale, these bots could be a smoke screen that manufactures a new privacy
try automatically tracking my web activity when for every 1 real one there are 100s of bots ...
posted by idiopath at 5:04 PM on May 1, 2012 [6 favorites]
try automatically tracking my web activity when for every 1 real one there are 100s of bots ...
posted by idiopath at 5:04 PM on May 1, 2012 [6 favorites]
Fraxas: Siri impresses me a lot more than weavrs do - but I wasn't talking about Siri. Also, this is uploading? Give me a break. I love stuff like this but these guys are jokers compared to the people at IBM or DARPA; or even a lot of game developers. By all means let's be optimistic about the future of AI, but the real work is being done elsewhere.
posted by adrianhon at 5:08 PM on May 1, 2012 [1 favorite]
posted by adrianhon at 5:08 PM on May 1, 2012 [1 favorite]
If horse_ebooks isn't a weavr, it should become one.
posted by The otter lady at 5:18 PM on May 1, 2012 [3 favorites]
posted by The otter lady at 5:18 PM on May 1, 2012 [3 favorites]
As far as I can tell, Weavrs creator Philter Phactory is only a handful of folks. There are only two people on their GitHub account. Frankly, I think it's impressive how much they've done. Siri was an entire company before Apple bought them, backed with more than 23 million in VC money. Wolfram Alpha, which provides the 'answers' is 20+ million lines of Mathematica code, developed over many years by Wolfram Research which has more than 400 employees.
But it's true, my Weavr has a terrible end game.
posted by jeffkramer at 5:33 PM on May 1, 2012
But it's true, my Weavr has a terrible end game.
posted by jeffkramer at 5:33 PM on May 1, 2012
Is this what we called "agents" last century?
posted by scruss at 6:02 PM on May 1, 2012 [4 favorites]
posted by scruss at 6:02 PM on May 1, 2012 [4 favorites]
Having read Perdido Street Station a little while ago, I'm less than comfortable with calling them Weavrs. I'm afraid they're going to cut my ears off and make a necklace with them, or something.
posted by LMGM at 6:02 PM on May 1, 2012 [4 favorites]
posted by LMGM at 6:02 PM on May 1, 2012 [4 favorites]
So after all this innovation, it seems like we're FINALLY going to develop realistic AI... for more effective spamming.
posted by lubujackson at 6:26 PM on May 1, 2012 [3 favorites]
posted by lubujackson at 6:26 PM on May 1, 2012 [3 favorites]
When you can't solve a problem, redefine it to something you can solve. Instead of trying to make turing-test passing AI, redefine conversation to take place in 140 character snippets and binary 'like / don't like' decisions.
posted by Pyry at 6:44 PM on May 1, 2012 [8 favorites]
posted by Pyry at 6:44 PM on May 1, 2012 [8 favorites]
By all means let's be optimistic about the future of AI, but the real work is being done elsewhere.
To be perfectly honest, I DO think there's something intriguing to this webbot approach to AI, almost as an afterthought. It goes to the fundamental purpose of AI - to have a program act in a way indistinguishable from an intelligent being. There's the top down approach of trying to improve the simulated (or generated) intelligence of software OR there's this Eliza/weavr approach of simply mimicking a human's interactions with a computer.
Part of what makes this fascinating to me is that it is a solvable problem, incrementally improving, with a very big payoff (probably by spam, mostly). It seems likely that software capable of passing the Turing Test would appear first through this sort of indirect approach rather than from a head-on tactic, because humans act in a psuedo-random way that is more easily patterned than replicated.
Step by step, weavrs are going to get more complicated and more robust. Tying together services and following real trends, following the swaths of data humans have been leaving in their wake these last few decades, imitating it - people are already following some bots on Twitter, many not knowing that they are reading bots. What happens when people start trusting bots and listen to their (generated) concerns? Or voting for a politician because everyone online (bots) was posting positive things on their blog and twitter feeds, even people (bots) who normally support the other party?
That bright future, closer than you think!
posted by lubujackson at 6:49 PM on May 1, 2012
To be perfectly honest, I DO think there's something intriguing to this webbot approach to AI, almost as an afterthought. It goes to the fundamental purpose of AI - to have a program act in a way indistinguishable from an intelligent being. There's the top down approach of trying to improve the simulated (or generated) intelligence of software OR there's this Eliza/weavr approach of simply mimicking a human's interactions with a computer.
Part of what makes this fascinating to me is that it is a solvable problem, incrementally improving, with a very big payoff (probably by spam, mostly). It seems likely that software capable of passing the Turing Test would appear first through this sort of indirect approach rather than from a head-on tactic, because humans act in a psuedo-random way that is more easily patterned than replicated.
Step by step, weavrs are going to get more complicated and more robust. Tying together services and following real trends, following the swaths of data humans have been leaving in their wake these last few decades, imitating it - people are already following some bots on Twitter, many not knowing that they are reading bots. What happens when people start trusting bots and listen to their (generated) concerns? Or voting for a politician because everyone online (bots) was posting positive things on their blog and twitter feeds, even people (bots) who normally support the other party?
That bright future, closer than you think!
posted by lubujackson at 6:49 PM on May 1, 2012
There is no weavr for the human spirit.
posted by steinsaltz at 7:15 PM on May 1, 2012
posted by steinsaltz at 7:15 PM on May 1, 2012
Is this what we called "agents" last century?
Yes, I believe so. Seems like it could be a good/useful implementation of the agent concept. Viewed from that perspective, this seems like an easy way to set up and play with a location-aware, advanced search notification thingy. If these catch on and stick around, it'll be interesting to see how they try to improve the signal/noise ratio and the arms race that'll develop with the spammers.
But to view this as anything more than glorified search is questionable. The angle that these are intelligent virtual beings living in and moving around the internet is just a silly distraction. This is no way near "uploading" one's mind to a computer.
AI will always be in the future, because once it works and we can see it's just a bunch of interconnected scripts, it's not AI anymore.
I think that's only true to a point. Once AI can join in and have insights into conversations like this, I don't think any one will even debate whether or not it is really "AI" (or "really AI") any more. That's what I think Turing meant by his Turing Test.
posted by Bort at 7:28 PM on May 1, 2012
Yes, I believe so. Seems like it could be a good/useful implementation of the agent concept. Viewed from that perspective, this seems like an easy way to set up and play with a location-aware, advanced search notification thingy. If these catch on and stick around, it'll be interesting to see how they try to improve the signal/noise ratio and the arms race that'll develop with the spammers.
But to view this as anything more than glorified search is questionable. The angle that these are intelligent virtual beings living in and moving around the internet is just a silly distraction. This is no way near "uploading" one's mind to a computer.
AI will always be in the future, because once it works and we can see it's just a bunch of interconnected scripts, it's not AI anymore.
I think that's only true to a point. Once AI can join in and have insights into conversations like this, I don't think any one will even debate whether or not it is really "AI" (or "really AI") any more. That's what I think Turing meant by his Turing Test.
posted by Bort at 7:28 PM on May 1, 2012
Peter Watts in his novel Blindsight had a good depiction of the Chinese Room theory.
posted by scalefree at 7:59 PM on May 1, 2012
posted by scalefree at 7:59 PM on May 1, 2012
Yeah I fear when business gets ahold of this, it's a whole new level of spam they could inflict into our eyes. Dammit, there's times when my eyes are completely closed & not I'm not ingesting any advertisements at all!
posted by scalefree at 8:05 PM on May 1, 2012
posted by scalefree at 8:05 PM on May 1, 2012
I'm going to call my Weavr "Randy" and his interests will be survivalism, guns, cabins, separatism, apocalypse, and aryan nations.
posted by symbioid at 8:32 PM on May 1, 2012 [1 favorite]
posted by symbioid at 8:32 PM on May 1, 2012 [1 favorite]
This scares me somewhat.
posted by robotot at 3:32 AM on May 2, 2012 [1 favorite]
posted by robotot at 3:32 AM on May 2, 2012 [1 favorite]
Hmm, yeah well... Randy makes no sense, there's nothing even close to his interests. It's like it just randomly scrapes various popular blogs and reposts random shit from them. Not cool.
posted by symbioid at 11:25 AM on May 4, 2012
posted by symbioid at 11:25 AM on May 4, 2012
« Older "It's OK to eat cooked meat that's been glued." | The 99% of the Web? Newer »
This thread has been archived and is closed to new comments
posted by jeffkramer at 3:30 PM on May 1, 2012 [2 favorites]