So Tuesday afternoon I was half-watching a stock ticker widget I've had pinned in a browser tab for no good reason (I don't even trade, I just like watching numbers move) and the Dow just dropped off a cliff. Like, straight down. A hundred and forty five points in about two minutes. And I remember thinking, huh, did something actually happen, or is this just algorithms being algorithms again.
Turns out it was neither, exactly. It was a tweet.
The AP's Twitter account got hacked sometime around 1pm Eastern on the 23rd and whoever had control of it posted something like "Breaking: Two Explosions in the White House and Barack Obama is injured." Total fabrication. The Syrian Electronic Army took credit for it afterward, said they'd phished a couple of AP staffers to get in. AP caught it and killed the tweet within a few minutes, the White House press office said nothing had happened, and the market clawed most of its way back before the hour was even out. By some estimates something like $136 billion in market value evaporated and came back in the time it takes to make a sandwich.
What gets me isn't the hack itself. Twitter accounts get hacked constantly, that's basically background radiation of the internet at this point. What gets me is that high frequency trading systems apparently read that tweet, or read the reaction to that tweet, and started selling before any human being had a chance to go "wait, is this real." Nobody called the White House. Nobody checked a second source. A string of text from a verified account moved real money in real time, automatically, faster than anyone could fact check it.
I've read a couple of the recaps floating around today trying to explain exactly which trading systems reacted and how, and honestly some of it is still murky, nobody's fully confirmed whether it was straight algo panic-selling or just human traders seeing the number craterand piling on. Doesnt really matter for my purposes. The point stands either way: we built a financial system that treats a tweet as a data feed, and this week it showed its hand a little.
And look, I use Twitter for news constantly. It's genuinely become my first alert system for anything happening anywhere, faster than any news app, faster than TV. I'm not going to pretend I'm above it. But there's a difference between me refreshing a timeline and going "huh, weird, let me check three other places" versus a machine that has no capacity to be skeptical at all, just executing sell orders on pattern match. That's the part that should probably worry people more than it seems to.
Small aside, unrelated except in the sense that it also happened this week: the Galaxy S4 went on sale Saturday and a friend of mine stood in line at an AT&T store for it like it was 2007 and the first iPhone all over again. Came home two hours later mildly annoyed that the box was smaller than he expected and that the home button still isn't as good as he wanted it to be. I don't know why phone launches still pull people out of the house physically when you could just order the thing, but I guess there's something to the ritual of it. Not for me. I'll wait for the reviews and the inevitable software hiccups to get patched out.
Anyway. Back to the fake explosion tweet for a second, because I think it's the more interesting story of the two even if it got less coverage than the phone did. Nobody died, nothing blew up, the market recovered by 1:10pm like nothing happened. But for about five minutes on Tuesday, a compromised password somewhere turned into a nine figure swing in the value of publicly traded companies, and that seems like the kind of thing that should get more than a day of headlines before everyone moves on to the next thing. It probably won't be the last time either. Passwords get phished, accounts get "verified," and somewhere a server is watching the feed and doesn't know the difference between news and noise.