Wired's bullshit "The web is dead" graph | Hacker News Reader

Wired's bullshit "The web is dead" graph

boingboing.net

394 pointsyanw73 comments

Chris Anderson and Wired manipulate facts to forward their theories with whatever new theory and bandwagon they are jumping on.

The data cited in 'The Long Tail' was not accurate[1], and actually supported the counter-argument. 'Free' was similarly vague[2] and inaccurate[3].

The entire m.o of Anderson seems to be to take a concept or theory that sounds interesting and package it in laymans terms for a gullible public. It is presented as 'research' or a 'breakthrough' and most average readers suck it up, but these theories are usually far from.

To see them doing it again with this latest theory is no surprise. Anderson really doesn't let the truth get in the way of a good story, especially when there are magazines and books to sell, and high-priced speaking gigs to book.

[1] http://en.wikipedia.org/wiki/The_Long_Tail#Criticism

[2] http://compassioninpolitics.wordpress.com/2008/03/01/critici...

[3] http://www.newyorker.com/arts/critics/books/2009/07/06/09070...

tl;dr Chris Anderson is like a less convincing Malcom Gladwell.

Free also ran into some nasty plagiarism accusations: http://www.plagiarismtoday.com/2009/06/24/the-chris-anderson....

Part of the reason behind Squeezed Books (which I recently sold on) was to 'deflate' books like that. Why buy something that's just a fad and can be summed up in a page or two.

For some reason Squeezed Books sounds like a more modern version of Readers Digest. What could possibly be lost by reducing 'War and Peace' to a 20 page summary?

Squeezed Books is only for business books, which are inflated in the first place in many cases: they could be restated in a few pages if you take out all the case studies and repetition of the Key Points and so on and so forth. Even if it's a novel, good idea, you can't sell a 10 page book, so they have to add enough stuffing until it's big enough to sell as a book.

They should have some sort of equivalent App Store for ideas. YOu pay 5$ and a file is downloaded to your computer. You open it up and it says "Have a list of stuff to do, review it daily in the morning, act on it during the day. Be organized."

The big issue is that speaking tours would be pretty much out of the question.

In economics, ideas are an interesting category of product. They're fairly difficult to exclude people from - you can't "sell" an idea to 1000 people and not expect to see it spread, if it's good.

They're often quite non-rivalrous, too, although the right one at the right time might mean that one person gets the most benefit from it.

What could possibly be lost by reducing 'War and Peace' to a 20 page summary?

I suspect that nonfiction is often more palatable to summarization than fiction because its main purpose is to convey facts. I don't think it's a coincidence that so many books start as magazine articles but then don't quite have enough material to sustain 200 pages of information. This is part of my problem with The Shallows: http://jseliger.com/2010/06/28/the-shallows-what-the-interne... , Rapt: http://jseliger.com/2009/08/16/rapt-attention-and-the-focuse... and some other books.

I think some authors would prefer to get away from it, but the publishing market, even though it's slowly moving in that direction, isn't quite there yet. If you're shopping around a solid 75-page piece, it's too long for a magazine (or even most journals, if it's research), yet too short for a book. So you either chop it or pad it, usually.

A good compromise, I'd think, would be to divide the piece into modular components with a repetitive (or narrative) structure, and publish them serially as a sort of "miniseries" feature in a magazine. This is how many pieces of novella-length genre fiction were originally published, for instance.

Measuring "the web" based on amount of bandwidth used is kind of a ridiculous metric. Time spent engaging (or some such metric) with a particular internet service might be more relevant.

Video uses a seriously disproportionate amount of bandwidth per use.

Next up: Ants are extinct! Their proportion of total mass compared to elephants is almost zero.

I see you point but...

"Their proportion of total mass compared to elephants is almost zero."

...that...doesn't seem true.

Hold on, I'll make a graph to prove it to you.

A graph on relative usage was used to prove something absolute...and you offering a graph on relative mass to prove relative mass...so almost but not quite. I didn't see anyone dispute the accuracy of the graph. People balked at the notion that the graph is proof that the web is dead.

"On average, ants monopolize 15–20% of the terrestrial animal biomass"

http://www.pnas.org/content/97/26/14028.full

Elephants are almost extinct because their proportion of mass is almost zero compared to ants!

That would be funnier if elephants were not, in fact, perilously close to being extinct, especially in the wild.

This thread is like a demolition derby for metaphors.

Tangent interested should check out http://www.greenermedia.com/hec.html for more on HEC(human elephant conflict) and to read about their recent trip to Sri Lanka(second) to finish filming a documentary called Common Ground. Can't wait to see it! Elephants and rural farmers need all the help they can get.

Also: Sri Lanka Wildlife Conservation Society www.slwcs.org

This is one of the things I love about HN. Someone makes a flippant comment and I get to learn something interesting cause someone else posts a response like it was a serious statement. I find myself correcting people's flippant comments often, or at least thinking about doing it, and most people just seem annoyed. I really enjoy learning something new even if it's tangential to the topic at hand.

There was a fascinating discussion on how geeks are not like other people in this regard:

http://news.ycombinator.com/item?id=1080272

This is one of the things I love about this site.

Right now, I'm half tempted to see if I can find something to submit or discuss that will lead to other tangents relevant to my interests.

It's funny, because I think I was interested in the Y-combinator itself when I came here...

That was StavrosK's whole point, he's obviously aware of it or he wouldn't have picked that particular example.

Give the guy some credit.

Thanks, it's a bit odd that people didn't get it from picking a ubiquitous, numerous species versus one that's almost extinct...

I actually had to laugh when reading it the first time because it seemed such a neat spoof on the article in so few words. You could have used bacteria or insects just the same. Just goes to show that there are many well known facts that turn out to be not that well known after all.

Perhaps your original intent would have been more immediately obvious if you had chosen a different animal that's also almost extinct, but not synonomous in people's minds with large size more so than the fact of their near extinction? Or was that also part of the original intent?

I admit that the analogy is a bit leaky, but the point is that ants are nowhere near extinct, even though they may be much smaller than elephants. Where it breaks down is if you consider the number of ants vs elephants globally. I probably shouldn't have said "total" mass, just relative mass.

Completely off-topic, but a quick google seach interestingly implies that ants hold a significant portion of animal biomas:

On average, ants monopolize 15–20% of the terrestrial animal biomass, and in tropical regions where ants are especially abundant, they monopolize 25%

http://www.pnas.org/content/97/26/14028.full

I didn't have as much luck finding a quick elephant biomass estimate. But giving a generous estimate of 1 million elephants, and the upper limit of the weigh of the larger African elephant at 26,000 pounds we can guess that elephant biomass is less than 26 billion pounds. Considering that's less than the biomass of humans in the US, I'd say ants probably actually do compare well to elephants in proportion of total biomass.

I apologize for any pedantry, but it was actually a question that piqued my curiosity.

I know, otherwise I wouldn't have picked an endangered species versus one that sends a billion members whenever you have a picnic! :p

Exactly. If I watched HD video on my iPhone, I would destroy my 2GB cap very quickly. But I can use the web and play QRANK and such constantly for a month and barely put a dent in it.

Not only this, but isn't video delivered over ... the web?

Well if you define the web as 'HTTP traffic on port 80' then... yes, a lot of it is. A lot is also delivered over RTMP (ie, streamed Flash video).

edit: RTMP not RTSP (that's Windows Media, Quicktime etc).

Which means that as streaming video moves from Flash to HTML5, web traffic (according to your definition) will actually increase.

If you define the Web instead as "stuff that happens in web browsers" then... you come to the same conclusion. What definition of the Web would not include all this video traffic?

I could see someone being interested in subsets, but yeah, they don't have a definition here. One that might be interesting is: what amount of time do people spend reading mainly textual documents online? That'd correspond somewhat to a classical view of the web as hypertext (you know, a "web" of interlinked documents).

That would, though, have to exclude not only watching videos, but also paying your credit-card bill on a bank's website, and other such web-as-thin-client rather than web-as-documents uses. Forums like this also occupy a weird place, since they're somewhat more like web-as-listserv/usenet-replacement than web-as-documents.

But, I'm not sure the end result would actually show a decline anyway. Wikipedia and blogs, for example, are more or less hypertext documents in the old sense of the term, just with some new conventions for organizing and editing them. Wikipedia alone probably rivals (exceeds?) the entire mid-1990s internet in size and readership.

This whole logic train is pretty much made of awesome.

I was laughing pretty hard by the "elephants are nearly extinct" comment.

Manipulated data sets are like bikinis - they show you what they want to show, and hide the bits that you want to see.

The way I heard it was that what they conceal is more important than what they reveal.

It seems to stem from the whole "A good sermon is like a skirt: long enough to cover the essentials, but short enough to maintain interest" (paraphrased). Not the same meaning, but it appears that the phrasing came from there.

The data set is not even manipulated. The graph in itself is not worthless, not wrong, not even uninteresting, it’s all the stuff around it (the interpretation) that’s bullshit.

You should probably never use such a skewed graph alone, though, without showing growth or decline. Relative shifts are genuinely easier to see with a graph like that (but internet traffic is probably not the right measure if you want to know something about cultural changes, time spent would probably be better; to wit, reading 1 MB is a heck of a task, listening to 1 MB won’t keep you busy for very long and blink and you will just have missed your 1 MB video).

Thank you for actually taking the time to make the graph we all instantly thought of when we saw that ridiculous wired article.

> Does 50MB of YouTube kitten represent more meaningful growth than a 5MB Wired feature?

I'd hate to have to answer that one with a clear consciousness. That particular article, probably not. But wired has had some excellent stuff in that past:

http://www.wired.com/wired/archive/4.12/ffglass_pr.html

I wished they did a lot more of that stuff than they do.

It seems like they're mostly focused on geek culture and various things that will get them hits/attention. I haven't really read Wired in the past, so I can't make a comparison to their previous habits, but that definitely seems to be their tone/focus. I subscribed to their RSS feed, but quit after a while because they simply had too much noise to signal. They do post quality stuff occasionally, but I prefer to just let that come to me rather than try to sift it out.

Does anyone remember ye olde times before 2000 when we all still had to rely on DNS? Good thing that clunky crap is long dead!

Video became part of the web experience long time ago. It can't be placed in a separate class unless it's IPTV or something that does not use a browser.

I don't understard why DNS was ever proportionally bigger than e-mail or Telnet.

In addition, if they wanted to say that more non-HTTP applications are on the net, they should point an increase in the "Other" class in the graph.

I guess they've choosen the wrong graph to support their article.

DNS was used to resolve hostnames for almost every other type of internet traffic, including email and telnet, even in the early '90s. As the size of data increased over the years, the size of DNS queries didn't change much, so its proportion of overall bandwidth usage quickly disappears from view in the graph.

Most of the video traffic is likely transferred over HTTP and accessed via web interfaces such as YouTube, so it's a bit disingenuous to distinguish video from web traffic - if you're going to make that distinction, why not put images in their own category as well?

#!/usr/bin/ruby

vapid_pronouncement = "#{dictionary.get_noun} is dead."

Why the downvotes? It's totally true. If the title includes "is dead" or "killed so-and-so" then it's going to be bullshit.

Because who puts ruby in /usr/bin?

  jam@jam-desktop:~$ which ruby
  /usr/bin/ruby

    #!/usr/bin/env ruby

    # you insensitive clod!

It's /usr/bin/ruby for the default preinstalled 1.8.7 on OS X.

Same for Ubuntu 8.04

I'll tell you who: People who aren't sexually attracted to donkeys.

Arch Linux does.

That's not really fair because everything goes into /usr/bin in arch, they don't even have /usr/local/bin in the path ( though I'm pretty sure they do still have /bin)

You should.

Who doesn't use package management? Nobody I want to have bits on systems I have anything to do with, that's who doesn't.

(It was a joke, I know it usually goes in /usr/bin/. That's where mine is)

I'd like to make a graph of the percentage of people that actually see a comment at negative reputation when there is a "why the downvotes" comment immediately following it.

So, by this metric shouldn't they be arguing that DNS died in the mid-90's, and that nobody uses it anymore?

You don't read Wired for the articles.

Wired is Maxim for people who can't get laid.

Maxim is for dumb guys who don't get Wired (or laid).

That too.

It's a pity because I usually find a needle or two of interesting ideas in Wired but having to sift through a haystack of hyperbole and poor logic is deeply unenjoyable.

It's not really acceptable on HN to editorialize the title like that. In fact, I believe it's explicitly forbidden.

I don't get why people are upvoting this story so much. Yep, Wired's wrong, but it's really not very interesting.

My problem with the graphic, beyond the proportional vs. accurate argument, is that it's misleading even in the context they show it. Outside of things like Skype, P2P is not an active-use process. Streaming video is clearly something that people largely watch on the Web.

When I saw that article today, I just thought it was complete crap. It's simply alarmist, the magazine form of using scary headlines to try to sell newspapers.

I don't know about anyone else, but I generally need the web to find my videos. Same with my peer to peer downloads (assuming thats BitTorrent). If the web literally died, it would take everything else in that graph with it.

Assuming web traffic means HTTP, it's curious that it's not zero in 1990.

LOL, Wired likes to predict the future - http://news.ycombinator.com/item?id=1612808

Reply on news.ycombinator.com