Slashdot Mirror

The truth is, spammers will eventually get around any form of captcha, if it's important enough to them. As long as a human can decipher your captcha, there's always Mechanical Turk.

And they've already got a way to do so cheaply
https://www.mturk.com/mturk/we...

Isn't this pretty much what topcoder already does?

Without having read the actual PDF (naturally), the description basically matches the competitive software component marketplace that topcoder provides. And, yes, like other commenters have mentioned, it seems like a Mechanical Turk style race to the bottom. Thankfully their component model seems a litte broken to me, so, perhaps that's why it hadn't taken over the industry.

Yeah, I honestly thought this was what mechanical turk was for....

Amazon's Mechanical Turk
Netflix Tagger

There's a couple that didn't exist a few decades ago, which are currently in the 'too hard to automate' category. They might not remain there forever, but these are arguably not skilled jobs, just normal information processing jobs that most desk jockeys would do well at.

How are the fees calculated?: "Amazon Mechanical Turk collects a 10% commission on top of the reward amount you set for Workers. For example, if a HIT reward is set to $0.20, Amazon Mechanical Turk collects $0.02 for each assignment." So, the worker gets 91% of the total amount paid ($0.20/($0.20+$0.02).

Well Amazon has something similar as a service: Mechanical Turk

Besides, machine learning is rather widely used today, but you need something for them to learn from.

I wonder how hard it would be to feed the input into an existing game engine? Gamers could identify potential targets and based on reputation / number of ids the target could be investigated further. You could use something like Amzon mechanical Turk to set challenges.

With the Army's Blimp providing more data, analysis will become increasingly more challenging.

The special cases that you describe (similar files with slight variations) are much difficult to handle programmatically. If you expect the number of such files to be small, then I would just handle them manually after the rest of the dedupe process is done. However, if you think there would be numerous such files and would require a non-trivial amount of time to classify, then I would consider automating the step using a service such as Mechanical Turk from Amazon. With MTurk some real person is involved in the classification loop (I don't recall what they charge, but it's pennies for each classification request).

[SARCASM]Where can I sign up[/SARCASM]

Try this first, you may earn your extra $2000/year easier and earlier.

If you want an army of people to shill for you at various online communities, you can just head on to Mechanical Turk. There you literally pay pennies to people all around the world to do a manual task you specify, like shilling.

You can sign up here. No guarantees you will always be schilling Slashdot, and the pay isn't really that great. Also, you'll apparently spend more of your time schilling if you sign up on this one. The barrier to entry is high for a non-Chinese native, though.

So there could be no good jobs if they charged $10 for the cable?
Somehow amazon can sell cables at a fair price, I bet they have some good jobs to offer as well.

https://www.mturk.com/mturk/welcome
Enjoy your future profession.

That data is from two months back, before Google Places appeared in web search. Now, it's worse. There's a whole mini-industry in the "black hat" search engine "optimization" community creating phony Google Places entries. Here's an ad on Mechanical Turk today:

Reno Gym - Google Maps Promotion (Client QMDHKOB)
Requester: Smartsheet.com Clients
HIT Expiration Date: Dec 18, 2010 (10 hours 52 minutes) Time Allotted: 60 minutes
Reward: $0.25 HITs Available: 2
Description:

• Follow Instructions on PDF attached for BUSINESS ADDRESS (1)
• Repeat Instructions on page 5 to 14 for BUSINESS ADDRESS (2) and (3) below.
  GMAIL ADDRESS: [Create a new Gmail Account] PASSWORD:
  BUSINESS ADDRESSES:
  • (1) 6370 Mae Anne Avenue, Reno, NV 89523
  • (2) 4784 Caughlin Parkway, Reno, NV 89519
  • (3) 18603 Wedge Parkway, Reno, NV 89511

  BUSINESS TITLE AND FULL ADDRESSES:

  • (1) Anytime Fitness 6370 Mae Anne Avenue, Reno, NV 89523 (775) 746-8400
  • (2) Anytime Fitness 4784 Caughlin Parkway, Reno, NV 89519 (775) 622-8034
  • (3) Anytime Fitness 18603 Wedge Parkway, Reno, NV 89511 (775) 852-7007

  WEBSITE URL: http://renogyms.org/
  GOOGLE PLACES URLs:

  • (1) http://bit.ly/cERse6
  • (2) http://bit.ly/9u33HG
  • (3) http://bit.ly/aKM8CI

Keywords: Smartsheet, Reno, Gym, Google, Maps, Promotion, QMDHKOB

Google Places spamming hasn't been fully automated yet, so we get to watch spammers outsource their manual spamming. Spamming Google Places is incredibly easy, much easier than creating the link farms required to spam Google's old web search. See the instructions in "Dominating Google Maps- The Most Effective Spam Ever And What You Can Learn From It".

Google Places has been 0wned.

I'm rather surprised at how intelligent the spambots are becoming.

I'm guessing those are paid spammers from places like India or even the United States. Amazon's Mechanical Turk are full of shady jobs like this, and I'm sure they're not the only operation around. Pay somebody 10 cents to put a post on a forum, multiply by a thousand times, and you've got a really good search engine optimization for $100.

This might be cheap for actual human transcription: https://www.mturk.com/mturk/welcome

the system was built thanks to a database of millions of human-labeled images put together by Chinese workers.

I spent a brief amount of time checking out Amazon's Mechanical Turk, and this was one of the activities they offered pennies on the hour for. Yay for crowd-sourced globalization! 100 years from now, when many of the mundanities of life are automated, is this what minimum wage workers will be doing?

Instead of advertising, why not find someone who needs a mechanical turk? For example, why not require that someone accessing the proxy perform some work items instead of interacting with an ad? These work items might be more valuable then an ad impression, and thus could pay more.

Already been done.

Agreed! See also: A9 and Mechanical Turk. Amazon's been doing a lot of interesting projects.

Regrettably, this particular application of the concept is (IMHO) flawed. It is hard to argue that humans are more adept than machines for solving a problem like SAT (at least manually) and as many have pointed out, the dimensionality of the space is too grand for a suitable visualization.

In the space of VLSI design, higher-level problems grounded in physical space (such as macro floorplanning and large-block placement) would be much more amenable to this type of game.

You're right, it needs to be done by humans to be sure.

Amazon's Mechanical Turk should do the trick.

https://www.mturk.com/mturk/welcome

...it will simply require a human-level brain.

How about Amazon's Mechanical Turk service?
https://www.mturk.com/

This seems like the most elaborate slashertisement (for http://www.mturk.com/ ) ever posted.

Is there a reason that they can't just use Amazon Mechanical Turk for this?

This is a perfect project for Amazon's Mechanical Turk -- http://mturk.com/

They employ Mechanical Turk?

I don't know... That's just what came to mind when I read "cloud staffing".

This can be brute-forced. Are you going to type simple question-answer pairs?

And if you start using patterns, so will the spammers -- "third" and "fourty-second" can be parsed easily enough.

It's also vulnerable to Mechanical Turk and similar attacks.

Fairly easy to implement? Sure. But trivial to overcome, if you're large enough for them to attack you directly.

You actually have a shot if you're small, though.

We just rolled out something simple -- I think it was even a FOSS library -- which sends some sort of challenge in JavaScript. Someone would either have to be automating a real web browser, or targeting our site specifically -- which might eventually force them to at least run a JavaScript engine.

That pretty much killed our comment spam overnight.

Obviously, it can't last -- as I said before, they could use a real browser (or use Mechanical Turk and use real people) -- or they could specifically target our platform (SpiderMonkey would pretty much take care of it).

You know, maybe I am a few links shy on my paperclip necklace, but don't you think it kind of conspicuous that while said DoS attack is going on, this submitter not only informed Slashdot about it, but actually pointed us all to the still-left-standing mirrors... as if to try and trigger the Slashdot effect.

An alternative way to game back is to contest every patent granted as soon as the grant comes out by claiming the "Obviousness" and "Prior art" statements. Just make a short use of Google and most patents will fall short. At least that may make the patent office saturated enough to keep their heads down.

One alternative is to use the Amazon Mechanical Turk to get help to hunt for stupid patents. Just raise some money first.

In case anyone hasn't put two and two together*, Amazon's Mechanical Turk is named in reference to the chess playing Turk from the article. Amazon's FAQ has more info.

* 5, for large values of two.

If it's going to cost 1 cent per request, why not just pass it to a real person on http://www.mturk.com/ and pay them a penny? You'll get a much higher accuracy!

You can always pickup a jiggit at http://www.thesheepmarket.com/.

About : http://users.design.ucla.edu/~akoblin/work/thesheepmarket/
Created with : http://www.processing.org/, http://www.mturk.com/mturk/

There goes Amazon's Mechanical Turk program. Time to get a real job.

Amazon S3 & EC2 are revolutionary, but at some point, it's a reasonable next step. The only big drawback of EC2 is the lack of persistence so it's hard to host a dataserver on there.

But the truly revolutionary service is Mturk. It's about packetizing tasks for humans! not for computers.

I'm going to donate some of my time to help finding him. The ability to help someone over the Internet is wonderful. However, the Amazon's mechanical Turk page (this one) refers to non-standard units like feet. I'm told I'll have to set the height on my GoogleEarth client at 1500 feet and that the aeroplane is 22 feet long, but this is like talking nonsense to me, as I can understand only metres (I'm from EU). I had to convert this, but I can't stop thinking that the page author makes the life of non-US volunteers a bit more difficult. Shouldn't important emergency-related webpages contain both the customary/local (feet) and standard (metres) units? Not that in this occasion this is of much importance (I just typed '1500 feet in metres' in Google), but I just wanted to raise this issue. What if you are in a life-or-death emergency situation and your ability to save someone depends on you understanding a non-standard unit that someone thought everyone would know about but in reality it is used only in some parts of the world out of custom? This problem of course also exists in many non-emergency situations... eg last time I got to England whenever I was asking locals where I could find a shop etc I had to deal with yards, and I had to get our my PDA to find out how far I should walk. With the global commerce and increasingly international communications, can't all countries agree to a single set of universal units and fund campaigns to help common people get accustomed to the standard units? Apart from language barriers, different sets of units really make intercultural communication harder than it needs to be.

I had the same problem at first in firefox and IE in windows. The FAQ below answers it. Short answer, click accept HIT at the top of the screen to log in with your amazon account first or it will not enable the forms. The "reward" discussed below is how much a site may pay you per recognition task, this particular task pays nothing.

http://www.mturk.com/mturk/help?helpPage=gettingst arted#what_HIT

How do I get started?

You can get started right away exploring Amazon Mechanical Turk, and finding work you want to complete. When you accept your first HIT, you will be prompted to sign in with your Amazon.com account e-mail address and password. If you already have an Amazon.com account, you can simply sign in. If you don't yet have an Amazon.com account, you can easily create one on the spot.

What is a HIT?

A Human Intelligence Task, or HIT, is a question that needs an answer. A HIT represents a single, self-contained task that you can perform to completion and collect a reward. You may do as many HITs as you like, whenever you like, in any order.

It's the Mechanical Turk approach. Amazon is doing it.

This project isn't the first of its sort: Amazon has the Mechanical Turk project, where users perform various tasks similar to CAPTCHAs for amazon.com credit.

http://www.mturk.com/

Also, Amazon has a pretty cool program where you can perform HITs (Human Intelligence Tasks) for a few cents each. They have a lot of stuff like transcribing podcasts, identifying stuff in satellite images, etc.

Anytime you can take a problem that is hard for a computer and easy for a human and make a successful game out of it you have a winner. I for one wish them the best of luck. check out google image labeler for another example of this. Amazon with their mechanical Turk went in the other direction, trying to pay you 1 cent (or something small) for each thing you. Check out this entertaining video on Human computation presented by the guy who originally came up with the google image game.

If they really wanted good captchas, they need to start using problems that are very easy for humans to solve, but very hard for computers to solve. For example, picking the one photo of a puppy out of a matrix of photos of full-grown dogs.

Image identification raises it's own set of problems. If you're working with photos, what is your source of photos? You're going to need a lot, if you've only got a thousand or so images, spammers will scrape them from your site and flag them by hand (possibly outsourcing the work, say through Amazon's Mechanical Turk). If you use a shared resource so you get get mind boggling numbers of photographs ("Bob's Puppy Captcha Service") you just create more incentive for attackers to index all of the service's images. (Yes, the service will try to make it hard, but armed with a botnet you can make large numbers of requests every day by visiting actual sites using the authentication system.) You won't want to go with free, publicly available images (say, pulling from Flickr's Creative Commons licensed images, since an attacker can easily scrape those images and pull the tags and description to automatically generate identification information. The only practical way to limit access to the database is to charge for it with a price high enough to discourage spammers, suddenly making the proposal fairly expensive.

Once you've got your system, you have to assume spammers are building up counter-databases that identify images they've seen before. You could try to fight back with simple distortions of the image to make them harder for a computer to identify, but you can't do much without making it hard for humans. And while having a computer identify objects in photographs is hard, having a computer identify highly similar images is pretty easy.

Ultimately a spammer or similar attacker interested in defeating Captchas doesn't need a 100% success rate. 50% is probably plenty; they're typically going for volume. Assuming a typical "three tries before we block your IP for a while or the exponential back off gets too long" system you can get a 50% success rate on those three attempts if you have a mere 21% success rate on each attempt. If the attacker wants a 90% success rate over those three tries, you only need a 55% success rate on each attempt.

You can see the type of work available for anyone to process on Amazon's Mechanical Turk right here: http://www.mturk.com/mturk/findhits?match=false

It's things like helping categorize images or finding specific things in databases of images or inspecting contracts -- you know the kind of stuff that's really easy for humans but is really difficult for computers.

I've tried a few in the past, however, most of the available "HITs" pay only a few pennies a piece, so I'm not about to go quitting my day job to sit at home fulfilling these requests quite yet.

This is clearly a patent for Mechanical Turk, an Amazon service that's been mentioned many times on Slashdot in the past. Whether or not this is something patentable should be up for debate, but what's so scary about it?

Perplexed as to why Amazon would launch two separate Question and Answer services, NowNow and AskVille I did some googling and thought I'd share what I came up with.

Firstly, it turns out that someone has already asked this on AskVille and the answers are fairly to the point. Also O'Reilly Radar has a post about the two services.

The gist of it is: AskVille is like Yahoo! Answers and NowNow is like 82ASK. NowNow is specifically set up for mobile users who need to find answers quickly. Questions are farmed out to Mechanical Turk where people are paid real cash money to answer questions.

However, Amazon have completely dropped the ball with NowNow:

1. The interface is mobile email! It's specifically for mobile phones and yet they've used email instead of the obvious choice of SMS.
2. Because they've chosen email instead of SMS they've also made the billing procedure more complicated. While the service is beta it is free, but I presume that in order to use it when it goes live you will have to set up some sort of Amazon account. So you have to know in advance that you might one day need to use the service, unlike with 82ASK.
3. Mechanical Turk is notoriously badly paid, and absolutely anyone can join it. So the researchers aren't vetted or trained like 82ASK's texperts are, and they're nowhere near as well paid as texperts, so they can't guarantee the same quality of answers as 82ASK.

AskVille has some interesting twists on Yahoo! Answers. Like an online game answers earn virtual money, and there is more community vetting and rating of answers and answerers then there is on Yahoo! Answers.

If they'd been really innovative they would have merged the best bits of each into on single product with both mobile and web access. If they'd done that there would be reason for 82ASK to be worried. But as it is I think they've blown it.

Jim Gray was closely associated with a number of satellite and aerial imaging projects. A group of his friends have been putting together a set of satellite images covering the area where he disappeared. They are looking for people to help sift through the images in the hope of spotting the boat. If you're interested in helping, Amazon have set up a site here:

http://www.mturk.com/mturk/preview?groupId=J0XZ58S TDWJZ5QY4F9M0

Why not offer your help to find Jim Gray? Visit http://www.mturk.com/mturk/preview?groupId=J0XZ58S TDWJZ5QY4F9M0

For those who want more background info of this search effort, I am posting an email from Werner Vogels (without his permission):

Through a major effort by many people we were able to have the Digital Globe satellite make a run over the area on Thursday morning and have the data made available publicly. We have split these images into smaller tiles that can be easily scanned visually and stored into the Amazon S3 storage service. We then created tasks for reviewing these images and loaded then into the Amazon Mechanical Turk Service.

This is where you come in. We need your help in reviewing these images to see whether you can locate Jim's boat in any of these images. Please go to the Amazon Mechanical Turk site and help us find Jim Gray.

The weather conditions were not ideal as some areas were cloudy, but we can still look for him in those places where there is a somewhat clear view. We hope to get more satellite data in the coming days of a wider area. The current images are panchromatic with a 0.82m, and Jim boat would be about 6 pixels in size. Please visit the Amazon Mechanical Turk site (http://www.mturk.com/mturk/preview?groupId=J0XZ58 STDWJZ5QY4F9M0) for more details.

I have to stress that many individuals and companies are to thank for making this possible; many academics friends relentlessly worked around the clock to get access to the data, many industry friends of Jim functioned as connectors to hook up officials and individuals, and people from NASA, Digital Globe, Microsoft, Google, Oracle, Amazon and others worked hard get to the data collected and available on a very short time scale. The Mechanical Turk team worked deep into the night to make this work.

Now it is your turn, go find Jim Gray.

Amazon has been doing this (perhaps in conjunction with others) for over a year now. The Mechanical Turk was primarily tested for identifying whether street-level images pointed at the proper address.

I assumed this was to make it possible to see what your destination would look like after creating map directions....

Wouldn't it be nice to see on your GPS direction screen both the street-level map, and a picture of the corner you should turn at?