Hacker News Reader

Comment by shawnz | Hacker News Reader

Were you using an unusual user agent or a VPN or something?

Twist: he is a robot, and when he sent them a photo, it was a robot photo. Square metal head, lights for eyes, antenna coming out the top. The works.

Maybe he isn't a robot, but his given birth name was zxcvbn4038

or bobby tables

"Little Bobby Tables we call him..."

That is my favourite XKCD comic!

A friend of mine uses a fake last name on Facebook because her real name is Fake, which they refuse to accept as real.

(Her name is actually Faux, which means fake in French.)

That would have been hilarious! Somebody should try it when they I accuse you of being a robot.

I would assume a submitted photo of a robot would go over just as well as when I was once asked to add a picture to the company's internal website of employees by our HR dept. I never add photos of me to anything, so instead, in good humor, downloaded a random anonymous silhouette type avatar. That didn't work for them, and I was told "I need a picture of you". So, I open up photoshop, found a generic picture frame online, and added the word YOU inside the frame. Shortly after this, my supervisor comes in to pay me a visit. Some people just have no sense of humor.

Have someone take a portrait photo of the back of your head.

Or take the photo using a potato. The security at my gym requires a photo id - but you can barely see who's on the picture. It looks to be printed with a dot printer.

In much the same situation, I just said NO. A protracted argument unfolded over the next few days until the matter was settled by me skimming a copy of the Conditions Of Service booklet across the desk and saying, "Show me where it says I have to." Immediate manager was not happy, but it soon blew over.

I don't a real human looks at the photo. It will be CV process that tries to identify the image with whatever word proof was asked to be present.

Hence the high rate of rejections

I got banned from Twitter within 5 Minutes of trying to figure out the user interface, and my best guess as to why it happened is that their anomaly detection is overfit to their existing users. So when a new user starts exploring randomly instead of directly going to look at ads, they're behaving much more like a scraper than what the system has learned to recognize as a normal user.

I think that's very unlikely. There are over 100,000 new accounts created on Twitter every day so we would expect a large and frequent amount of bans if what you are saying were true. Furthermore most Twitter visitors don't have accounts so it would be pretty foolish to base their anomaly detection on a metric that most of their visitors fail.

It's actually well known that you'll get blocked by Twitter within the first 2 minutes of registering, whatever you'll do.

They probably use it as a way to enforce their "soft" requirement of a phone number. After verifying that you have a working phone number, you'll get unblocked.

Also, Twitter routinely renders their tweets unaccessible to non-logged-in users, due to rate limits on their api (which is used by their web frontend).

So, I'm not one bit surprised if anyone would describe Twitter as foolish: that's exactly how they act

So basically if the phone companies lowered the price of buying number lists then a bunch of web2.0 companies' valuations would evaporate?

I had Microsoft do this. They notified me that my account was being suspicious, so I needed to provide a phone number to use it. I had literally just made the account though, zero activity at all. It was just to force me to give a phone number to them for their data harvest

I ran into this with MS as well. However, they explicitly stated that it did not need to be my phone number. So since it was for a work thing, I ask to use my boss' number to get around the lockout. Seemed pretty pointless.

One of my land-line numbers is so spam filled, i ignore it and use it as the default phone number for such antics.

I disconnected the ringer so it doesn't bother me.

Welp, it's a good thing these companies don't have a large amount of control over a lot of people's lives. And luckily their business practices are transparent and we all agree they're fair.

This is fine.

Hey, I have some experience creating intentionally suspicious Twitter accounts.

There is no phone number requirement. They're just as happy with an email address.

If you provide neither of those, then your account functionality is restricted (with a big obvious banner showing at the top of every page), but you still don't get blocked.

Once I needed to create several accounts at once and then have all but one of them follow the last one. Those accounts got blocked, but it's hard to blame Twitter for that.

> There are over 100,000 new accounts created on Twitter every day so we would expect a large and frequent amount of bans if what you are saying were true.

That's in fact true. Every single one of my Twitter account that I've created in the last few years have got banned within a few minutes even before I've had a chance to post anything meaningful with the account.

I wouldn't be surprised if the anomaly detection is overfit to banned users: comes from Kenya, or Russia, or China. Doesn't initially seed the account with in-person connections. Shares an IPv4 address or telephone number with other folks who have been banned before.

Why would they bother checking the user agent when it's the easiest thing to spoof (really, you probably can't even call it spoofing)

I'd bet that the vast majority of botters/phishers which Facebook deals with are amateurs. That said, I meant using unusual user agent software, not really just having an unusual user agent string.

No need for software, one can easy change it on the about:config of Firefox and/or there are some cool add-ons that can rotate/change the user agent of a Firefox on the "desired frequency"

We have a Wordpress site that suddenly saw a surge of hacking attempts. Turns out they all had IE7 user agents. Blocked that, not a problem anymore. Wouldn't surprise me if big sites use simple measures for such simple attack attempts as well.

Reply on news.ycombinator.com