Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

This is the text of the message I received (once for each account, all created back around the same time January 2007):

  Hi, dewitt

  Twitter believes that your account may have been compromised by a website or service not associated with Twitter. We've reset your password to prevent others from accessing your account.

  You'll need to create a new password for your Twitter account. You can select a new password at this link: ***

  As always, you can also request a new password from our password-resend page: https://twitter.com/account/resend_password

  Please don't reuse your old password and be sure to choose a strong password (such as one with a combination of letters, numbers, and symbols).

  In general, be sure to:

  Always check that your browser's address bar is on a https://twitter.com website before entering your password. Phishing sites often look just like Twitter, so check the URL before entering your login information!
  Avoid using websites or services that promise to get you lots of followers. These sites have been known to send spam updates and damage user accounts.
  Review your approved connections on your Applications page at https://twitter.com/settings/applications. If you see any applications that you don't recognize, click the Revoke Access button.
  For more information, visit our help page for hacked or compromised accounts.

  The Twitter Team 
Best of luck to the security and support teams. Days like these are not fun at all.


And for people trying to puzzle out who was impacted, several these accounts (all with random strings for passwords, btw) were barely ever used at all, often not for several years. The only thing they had in common was their early creation date, and hence relatively low user ids. My guess is that the hackers simply scanned user ids starting from 1 and worked their way up.


I'm seeing a ton of people I know on twitter complain - I'm user 5511, and these are people who joined at the very beginning too - so your theory is indeed apparently correct. I got an email myself, and reset my already super complex password.


User 5,260 and an active, daily users. Got an email that my account was impacted.


I don't know what # user I am (how can I get this information?) but my twitter account was also created in 2007 (on the 8th of April says whendidyoujointwitter.com) and is still active. Also received the email.


The Twitter API exposes a user's ID. Some Twitter clients (like Tweetbot) show this information. You can use http://mytwitterid.com/ to find yours.


Or just view source on twitter.com (after logging in) and search for the first instance of "data-user-id"


3750 and ditto.


That would explain the high incidence among hackers, who are more often early adopters. I've been surprised by how many people I've heard of getting the email, including people in this comment thread and myself, considering only 250,000 emails were sent out of their couple hundred million accounts.


Only 2950 accounts with IDs from 1 to 6136 still exist, so there's been pretty much a 50% attrition rate at that level.


Yeah I received the email from Twitter, account created in April 2007 and haven't tweeted since 2010.


Are there big gaps in the early user ids? I'm 4145801 and received this message.


Twitter employee, here. At one point in time, auto_increment_increment was > 1 on the MySQL master for uid generation. This led to many holes in the uid range.


Yeah - I'd say there are big gaps. I'm 1577581 and got the email. You can check people's join dates here http://www.whendidyoujointwitter.com/


Thanks! I'm 793689, joined 25 February 2007, and got the e-mail.


I've been suspecting the same thing. Two of my accounts received the email, both created several years ago, while none of my newer accounts have been compromised.


any chance it was through an app that you allowed access?


No. I suspect their email template was out of sync with this particular incident.


That means https://api.twitter.com/1/users/show.xml?user_id=12 13 14 were hacked too. And those must have been accounts of interest if not of others' and the gentry.


This is bullshit. I just got this message, and until I signed into HN I had no idea if Twitter was hacked or if there was a problem on my end. Which would be alarming, because all of my passwords are 30+ random characters, and I never reuse passwords across websites. Fuck you, Twitter.


Just a guess here, but maybe to get the emails out to users fast they re-used an existing template that was intended for resets due to 3rd party incidents.


Seems to be that way.

I just got a second 'Twitter Password Reset' email with more explanatory information now, almost two and a half hours after the first.


Even if this is true, it is very sloppy for a company the size of Twitter.


def random_password(n):

  random_char = '1' # chosen by fair dice, guaranteed random

  return random_char * n

Just kidding. That stinks. I'm guessing your password was quite strong. Any idea how many bits of entropy it was?

It sounds like at this point Twitter sent out the email in parallel with trying to figure out how these compromises happen. Since they salt the hashed passwords, they don't know how complex your password was. Of course, you should still change your password. I changed my 80-bit Linkedin password after it was stolen.


Agree. I didn't know what was going on. I use 17 random char passwords too and the way the email is written makes it sound like I did something wrong.


What could Twitter have done better? What would you do if your users' accounts were compromised?


Where and how do you store these individual 30+ character passwords


I use lastpass, it's a great product.


Doesnt seem a great idea. When it gets hacked they get all your passwords.


LastPass encrypts all of the passwords client side. Assuming you use a strong enough passphrase it shouldn't matter if LastPass gets hacked.


Twitter uses bcrypt, so in theory this hack should also be nothing to worry about.


is there an alternative where this is not the case?


My own solution is to have two different passwords for everything - one for banking and credit cards, another for crap like twitter/linkedin. I haven't changed my passwords for years (no point really, as you're likely to have the breaking as soon as they get your password).

I think there are risks with all solutions to the password problem.


I'll add a datapoint here too. I also had my password reset.

Creation date: March 2007

User ID: 2,7xx,xxx


Surely these sorts of messages are prime candidates for opportunist phishing attacks?




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: