
The reason FB is doomed is simple. They floated the company and the customer is now the ad-men and not the users.
The annoyances are building. It will go the way of MySpaz.
Facebook's data scientists have schooled two Princeton students who published a paper predicting that the social network would lose 80 per cent of its users by 2017 and go extinct shortly afterwards. In a mock study Mike Develin, Lada Adamic, and Sean Taylor, of Facebook's data team, used the same methods as the original paper …
That would be MIT.
If memory serves its the main library that has a dome and one year, many, many moons ago, someone put a big nipple on it calling it the breast of knowledge.
Note that other objects include a working telephone booth. (Yes, you really have to go back to the 70's and early 80's to see some of these crazy hacks.
And in the fall of '82, MIT 1 vs. Harvard/Yale 0; (Half time prank where they inflated a big balloon filled with talcum powder so that it wouldn't stick....)
I thought the original article had a reasonably intelligent point, but trying to debunk it with the same methods was never going to work. Note: this is all about discrediting the report by attacking its methods than about proving that it is flawed for any other reason... it's almost an ad-hominem.
Precisely. If the method of a scientific paper is shown to be wrong, the paper is wrong, i.e. its conclusions are flawed or unfounded. Once you have shown a conclusion is unfounded, you can safely ignore it, because it has been demoted from conclusion based on scientific method to opinion. The opinion could be right, of course, but without scientific backing it is just that, an opinion.
Attacking poor methodology is good practice in science. Much as I hate to come to the defence of anything Facebook, the data scientists are correct in attacking the method
"Note: this is all about discrediting the report by attacking its methods than about proving that it is flawed for any other reason... it's almost an ad-hominem."
No it's not.
No hard science degree for you.
(Also: The contention that less internet searches for it = dying is clearly erroneous. The quite opposite "It's so popular that people know how to find it with out googling!" is just as valid a reason for the decrease in search.)
IMHO, the response by the Facebook lad will be of great didactic value in A-levels and first-semester courses in illustrating various basic points about research and statistics (choose your data sources properly, apply the right methods to your experiment, and of course, correlation vs. causation) in a fun way and by relating it to something everyone concerned will know about (Farcebook) and many/most will have experience with.
I feel Dr. Develin deserves credit for this, while recognising that this is probably going to taint somewhat the results of the Princeton guys' peer review. Oh well, it's all experience after all. :)
Some of the comments on that facebook page are pretty amusing too. Complete with obligatory XKCD references (605) and Internet Explorer market share vs murder rate graphic showing (with the aid of suitable manipulation of the Y axis) an undeniable correlation.
Of course, if Facebook turns out to be history by 2016[1], the Princeton side will have had the last laugh.
[1] Or if Princeton run out of students by 2021. Bit of a Pyrrhic victory in this case though. :-(
But, even if FB dies by 2016, there may not be a factual causal relation corresponding to or supporting Princeton's work. For example, if an earthquake destroys SanMaAlto, and concurrently FB's underground servers overheat in agony at the whoosh sound of MZ and company vanishing, those events would have nothing to do with bored, defecting, or overloaded users checking in less on their accounts and friends' walls. Princeton will have just taken a SWAG, and coincidentlally a prediction.
Awesome comment on the correlation between IE market share and corresponding rate of murders. So preposterous, it was one of my best laughs of the day....
Speaking as a climate scientist, thank you helping to show me the error of my ways. The fact that correlation does not imply causation - which you alluded to in your posting - came as a complete shock to me and those colleagues who have also read your comment.
We are erecting a statue in your honour, then we are all resigning and getting jobs stacking shelves in Poundland.
I thought the original Princeton piece was intended to be a bit tongue in cheek. I mean it's titled "Epidemiological modeling of online social network dynamics" and it was put out by the Department of Mechanical and Aerospace Engineering. If it had come from either a bio-science or computer science related department, I might have thought it was an attempt at a serious treatise but coming from the aero-mech boys it seemed pretty clear that they likely had a pint or two too many and decided to extract the urine out of a company owned by a former Harvard boy.
Damn you, Eddy... I've got cramps, from my most intense laughter of the day... "twin half-sister".... You just surpassed my laugher-response to ie-vs murder rate correlations laughter...
Damn, my head is pulsing.... I gotta get back to drawing before I stumble up something that keeps me from being able to breath steadily...
Twin half-siste......
"Oh, in that case 'listen up' is the twin half-sister of pipe down, shut up, kick back, put out, chill out, dig in, hunker down and a whole bunch more, all right errr, alright?"
Errr, not alright.
Remove the second word from the examples you gave (with the possible recent exception of chill) and you change the meaning completely or leave no meaning.
Remove the 'up' from 'listen up' and you have a perfectly good word meaning what it always meant, way before the first plonker stuck a spare 'up' on the end.
I'll see your 'errr' and raise you a 'Doh!'
There are a lot of the very, *very* hard of thinking, judging by the near total domination of reality TV including at least one entire channel.
Plenty enough to keep fb going unfortunately. If not, something just like it, another pyrrhic victory since that will be shit too.
And Sherlock came and went so quickly.
I think we have reached a tipping point where 6% of GDP on university education has stopped being merely wasteful and started to be actually harmful.
The number of low quality papers published by thousands of PHD students and there mediocre professors is drowning out real science.
I mean is it really necessary for a software salesbody or a payroll clerk (sorry HR consultant) to spend four years of their life studying something totally irrelevant?
It could be. Or at least it could be the same root cause in the brain functionality. I'm very mildly dyslexic. I've noticed that these days my brain can get so far ahead of where I'm typing that you'd have no idea what I've mangled. So I'm trying to be more careful.
... what this "study"/"counterpoint" really shows is that the goo-tards are becoming increasingly meaningless to folks who are actually looking for real, useful information.
Seriously. Think about it.
Facebook? Looked at it once. Found it useless. I have a RealLife[tm].
::shrugs::
Facebook geeks are still making a fundamental mistake with their (tongue in cheek) attempt to discredit the Princeton prediction about Facebook. The methodology as used by Princeton works mainly for studying the spread of infectious diseases or ideas and therefore makes the assumption that Facebook acts more like an idea than an institution and a such it would end up having the same kind of demise.
The error of the Facebook rebuttal is to liken Facebook to an institution like a world class university (what modesty!) and dismiss the notion Facebook would be akin to a virus or as vapid as an "idea". Facebook believes, naturally, that they are about real connections and services being provided.
The idea under scrutiny is probably the one of "being connected on-line to my friends, families and lovers" but is Facebook justified to think they can keep the idea like "Facebook is the best way to maintain these connections" alive? Not before they can become a distinguished institution, I'm afraid, something that cannot be replaced by a new collection of cool upstarts that easily.
Both are distressingly common in science, I'd estimate that in my field somewhere between 20-40% of all published articles are junk. In this particular case, considering that these two guys were PhD students, I think it's a case of Baby's First Publication.
That students publish junk isn't really their fault, they're learning and still don't really know what they're doing, the real fault lies with their supervisor, who should have supervised, coached and advised them, and with the reviewers, who should have rejected publication.
I mean sure extrapolating a trend isn't always going to work, but ask Wolfram Alpha for the number of Facebook members per world population.
http://www.wolframalpha.com/input/?i=facebook+members+per+world+population
If Facebook continues like this and we don't have a plague that kills mostly non-Facebook users, their growth will be consider to be considerably smaller than the growth of the population. Cutting down on fake accounts doesn't seem to help eithere.
If you want to see why Facebook is doomed to failure, just look at its predecessors. Compuserve, AOL and others.
The community will last until a point. And then people will leave out of boredom. There is also a diversity of sites that offer similar parts of the service, without the over abundance of ads.
As people leave and new people don't join, they will be forced to push more ads to make more revenue off their shrinking population. Its an inevitable death spiral.
How long it will take, is anyone's guess.