When I started reading this, I didn't realize it was fiction. When I got to the point where the protagonist left the end-date off a query and saw a spike, I thought it was going to be explained by falsified data and lead into a (real) accusation of providing fraudulent metrics to advertisers. It wouldn't be the first such accusation. But it turned out to be a science-fiction story with unexplained time travel in it. Oh well.
I did think the PIG abbreviation was too good to be true and decided that I didn't want to read it if it wasnt anything to do with real life internals
"Read the article FIRST and then come back to read the comments.
Spoiler alert. Post is fiction. Yadda yadda."
Still a good short read IMO.
However, in the actual story, there was no "black box" to query, the ex-employees wrote all the complicated queries themselves.
EDIT: They turned ~$150K into ~$2.8 MM USD over about three years, before being caught, mostly through options trades it seems.
But then on September 6, 1997, something quite extraordinary happened: the graph shot upwards, recording a sudden and massive shift in the number sequence as his machines around the world started reporting huge deviations from the norm. The day was of historic importance for another reason, too.
For it was the same day that an estimated one billion people around the world watched the funeral of Diana, Princess of Wales at Westminster Abbey.
it is still running, here is the data chart from monday http://noosphere.princeton.edu/data/eggsummary/2016/summary-...
'The "misappropriation theory" holds that a person commits fraud "in connection with" a securities transaction, and thereby violates § 10(b) and Rule 10b-5, when he misappropriates confidential information for securities trading purposes, in breach of a duty owed to the source of the information. ... Under this theory, a fiduciary's undisclosed, self-serving use of a principal's information to purchase or sell securities, in breach of a duty of loyalty and confidentiality, defrauds the principal of the exclusive use of that information. In lieu of premising liability on a fiduciary relationship between company insider and purchaser or seller of the company's stock, the misappropriation theory premises liability on a fiduciary-turned-trader's deception of those who entrusted him with access to confidential information.'
The employees stealing the information would have still been liable. But had they simply accepted payment from the hedge fund in Bitcoin and not revealed their true identities to the hedge fund, they would have been much harder to find and prosecute.
"People have asked me if this is insider trading and, you know, sure it is? (If the allegations are true, I mean.) This is not "classical" insider trading -- trading or tipping by an insider at Chipotle or whatever -- but rather "misappropriation" insider trading..."
* example: you post a trade but you don't intend to execute it but instead hope to fool the market, then you cancel the trade and switch to other direction. forgot what that's called exactly
What a bizarre piece. How long did it take everyone else to realize it was fiction? For me, I did not realize until the very end -- and even then I still wasn't sure. It could just as well have been written by some delusional non-technical employee at Facebook.
Longer than it would have taken if I'd never worked in analytics.
I doubt there is anyone who has ever worked on analytics software who doesn't have stories about articles with negative page views, impossible traffic spikes, or data from the future. (And in fact Facebook's APIs are some of the worst offenders for producing garbage data.)
Using Facebook internal data to trade in ways no one can match isn't far fetched. Makes me wonder if Facebook should pivot to be a hedge fund based on signals no one else has access too.
Interestingly, the data could be equally effective for building actuarial models and selling insurance to customers (I believe Google has stated an interest in this in the past).
Does anyone know what the SEC implications would be for a company using its proprietary user-generated data as trading signals? Is there any precedent?
Here's what Eric Shmidt had to say on the subject:"There are many, many things that Google could do, that we chose not to do. One day we had a conversation where we figured we could just try to predict the stock market. And then we decided it was illegal. So we stopped doing that."
Information that other people don't have access to isn't by definition insider information. It's just stuff others can't or haven't figured out.
Also, if you've ever analysed any traffic the graphs are clearly fake. Trends don't decay that quickly. And the normal levels would have waves due to seasonal changes, weekends and so on. Type "VW" into Google trends to see what I mean.
But once it was like, "hey, we're predicting the outcome of a sports game before it happened based on this" then it's like "ooh, this is a cool fake story which just took a turn towards the supernatural, I get it..."
I'll also eat my shoe if there's a single person with the legal name "Julie Rubicon" in the world.
It doesn't seem impossible. I mean I'm not having a ton of luck but there are references to some people that may legally have the name.
Have a big mouth? Put steak where your mouth is!
I checked my calendar--nope, not April 1... And kept reading.
"The last thing I want to do is write this down, but I’m doing it anyway, partially because people ought to know what’s happening with the things they post here, but mainly (like 99%) because of Julie Rubicon and the spike."
Oh, I've just discovered a world changing dark secret brewing in the heart of Facebook. I'd better tell the world! But first I must build some... suspense!
If it had turned out to be real, I'd thank the whistleblower right after I slapped them for being a twit.
wasn't yet sure but I started to pay attention, noticed extra "woolly" style of writing (not bad btw, just another marker), the bit where he said it was a neural network that accidentally started preducting the future sealed it.
(If there was really a PIG group in some company, I'd imagine that their work space would be known as the "PIG pen", not the "PIG pod".)
So they knew it wasn't about a new car model, but some negative event.
1 - https://en.wikipedia.org/wiki/Mr._Penumbra%27s_24-Hour_Books...
I felt like the book was mostly filler added after the popularity of the short story.
Popular enough to turn it into a book? :)
Well, maybe it was because I'd read the story before (and really enjoyed it) so I couldn't get the same enjoyment from the book.
Anyway, good work and keep writing.
I read way too much science fiction. Back to work now!
Kind of relieved though.
Once I finished, when they repeated the name Julie I scrolled up, came back here and boom, fiction. lol
'I spent the day haunting the roof of MPK20, feeling jangly and nervous, staring out across the bay.'
i thought it was about her confronting the ethical dilemma about whether she should rat out on her employer or not. it feels more likely you would accidentally disclose your fake data generator than your future predictor.
No matter what you query, you get a convenient spike right around your time of interest – that must be nice to show clients!
One of the things that immediately flashed it as fiction for me was that the graphs had all the same shape, which if you've ever looked at trend graphs you will see they might all have a similar outlier quality to them but they build and sustain in different ways.
At Blekko (a search engine company) we did an interesting study on query traffic to see if you could "predict" the "hallmark" holidays based on search queries. The idea was that holidays like Valentines Day come up, people start thinking about plans or gifts before that, could we advise an advertiser when the "peak" planning session was so that they could maximize the impact of their advertising spend by focusing it during the peak? And if so what sorts of queries were people making that indicated they were doing holiday planning? The results were mixed. For things like Valentines it was easy, flowers, chocolates, bed & breakfast reservations sort of rose out of the general query stream, St. Patrick's Day? Not so much. But the data peaks all had different shapes appropriate for different levels of impact (Christmas shopping really starts in August among the back to school traffic for the really prepared). So looking at (and for) "interest spikes" like the ones in the story had a bunch of different shapes, some with slow onset and rapid decline, some with rapid onset and rapid decline, and some which were like soft swells on a breezy afternoon at the beach.
That said, the dataset made possible by Facebook's chat stream would be even better for those sorts of investigations.
And if you are looking for a longer, slightly better, fictionalized account of Facebook, "The Circle" by David Eggers is a quick, engrossing read and quite hard to tear yourself away from once you begin.
That said, what's described could possibly be done with existing technology . Could Facebook accurately move forward statements like "next Monday is going to be massive for Volkswagen" over time you could weight private messages from people who work for regulators higher. For more public events like an apple launch they could predict a spike easily just by all the media mentions of the date
The only "mistake" is that the CONFIDENTIAL watermark is a lower z-order than the chart, which defeats the purpose of the "watermark."
(the variance on the graphs intuitively felt a bit low, but then again it is facebook)
Apart from the funny combination I just chime in to recommend the book here, I think most of you here would enjoy that.
"Psychohistory is the name of a fictional science in Isaac Asimov's Foundation universe, which combined history, psychology and mathematical statistics to create a (nearly) exact science of the behavior of very large populations of people" source: Wikia
Or it's not entirely fiction or perhaps even factious. I remember Rupert Sheldrake mention in one of his mindboggling talks that he wanted to investigate potential psi effects on a much larger scale and was in talks with Google. If there were 'results' pertaining to precognition (or AI enhanced precognition of the crowd) would the public get to know about it?