Hacker News Reader

Comment by hyko | Hacker News Reader

the average person “consumes about 34 gigabytes across varied devices each day” — some 100,000 words’ worth of information - seems like an odd and misleading statistic.

View on HN

Yeah, that's silly. 34 gb is 34 billion bytes, so if each byte is a character and the average word about 5 characters, 34 gb should be more like 6.8 billion words.

Also, most of that 34 gb is probably video anyway, in which case we may as well say the ancients consumed many gb of information per day in the form of looking at the world around them.

That number could be made arbitrarily high... an hour of 720p TV is 1280 * 720 * 3 * 60 * 3600 is 238GB

Granted my eye averages/summarized the pixels and my brain ignored a lot of the content (i.e. I don't notice the clock ticking in the corner while the hero is fighting the bad guy), yet still an argument could be made that that's the amount of data I'm exposed to while watching an hour HD movie through my HD display.

You are mistaking words for information. This message has few dozens words but much more information :)

Oh, no way. The information conveyed through human language is an abysmal drip compared to the actual number of letters used to spell it out.

Even without knowing anything semantic, compression algorithms can squeeze text down to much fewer bits. And both of our minds together have a shared dictionary of concepts and thoughts gzip doesn't have access to. The actual number of bits conveyed is rather low.

English contains around one or two bits of information per character, so actually they are basically correct. See http://www-isl.stanford.edu/~cover/papers/transIT/0413cove.p... and http://www.anthology.aclweb.org/J/J92/J92-1002.pdf.

Yea, to put it in somewhat manageable terms, War and Peace in plain text is about 3.2mb (Project Gutenberg). I know damn well I am not consuming anywhere near that much daily. Doubt anyone is on a regular basis.

If the 34gb is in terms of the "video" you intake, even just going outside and bird watching or sitting in a park... I guess? Sure? Still, misleading really.

It’s also demonstratively untrue.

100,000 words is, by itself almost certainly under a megabyte.

(Avg word length (let’s say 6) * number of words = number of bytes)

If you count -everything- then you’re basically insinuating that people on data caps are either consuming less (and because it’s an average, everyone else consumes much more) or outright claiming that you can quantify data that is consumed outside of a computational device.

Agree. Also, even watching videos for 24 hours it seems unlikely that you'd go through 34GB worth. Maybe the size of video files has gone way up since I was a movie downloader, but 12 2 hour movies is probably < 34GB..

At 1080p60 it's probably doable (heck, I wouldn't be surprised if I've beaten 34GB when sick), but it's still a rather absurd figure. Are they counting each site visited, including javascript and CSS, without caching, and every video loaded on Facebook, even if unwatched? I doubt even such a measurement would approach 34GB on average.

It really depends on where you measure it. A 4k BluRay is somewhere around 60-80GB for a single movie. But then when that is uncompressed and sent to the actual screens you are talking multiple gigabits a second. With 4k60 being somewhere in the 12 to 18 gbit range at the peaks. If it were at 16gbit for 2 hours, that would be ~14TB of information transferred.

If you are talking streaming bandwidth, Netflix says 25mbit is required for 4k video streaming. If it were to use all of that for 2 hours that is 22GB for a movie. Though its likely only using half that. I haven't actually measured it. And then there is the same inflation to what is sent to screen.

As a point of interest, there are audio- and cinephiles who prefer 10bit color depth, 5.1 FLAC audio. I'm not one of them, but I stumble across their files sometimes, and I have seen stuff like Blade Runner (~2h run time) to the tune of 20-30GB (IIRC).

10bit h.264 was actually introduced as a way to reduce file size. I'm not familiar with the details, but apparently it reduces noise in the compression algorithm.

You can even find 4K remuxes weighing in at over 50 GB.

a blu-ray rip with no reencode can easily be 20-30GB. even the higher bitrate 1080p encodes found on public trackers can be 8-12GB. pretty easy to go through 34GB of high quality 1080p movies on a domestic plane flight (if you have the storage).

seems it stems from a study called "How Much Information? 2009 Report on American Consumers"

from the management summary :

"In 2008, Americans consumed information for about 1.3 trillion hours, an average of almost 12 hours per day. Consumption totaled 3.6 zettabytes and 10,845 trillion words, corresponding to 100,500 words and 34 gigabytes for an average person on an average day. (...) We defined “information” as flows of data delivered to people and we measured the bytes, words, and hours of consumer information. Video sources (moving pictures) dominate bytes of information, with 1.3 zettabytes from television and approximately 2 zettabytes of computer games."

I found a copy here : http://group47.com/HMI_2009_ConsumerReport_Dec9_2009.pdf

The only way I could think of to interpret this, is to take the content in 1 hour of video and figure out how many words it takes to give you that same knowledge. This will vary based on the type of contents (nature documentary, vs a thriller move, etc). But to start with, assuming watching a regular drama. An 80,000 word novel can turn into a 2 hour movie, so each our of video is worth 40,000 words. But a lot is left out when making a novel into a movie, so you may need to take the Reader's Digest version, which gives you 20,000 words per hour. That means that people take in about 5 hours worth of video content a day.

Note, these numbers are my best guess, would like to see better numbers filled in.

Especially when we consider that the computer is the one consuming bytes. English's information entropy per character is quite low, & we consume text, not bytes

https://cs.fit.edu/~mmahoney/dissertation/entropy1.html

If I consumed that much per day on average I’d have major bandwidth overage charges at the end of the month. (In Canada, at least, where caps are often ~200-300GB)

The precise number doesn't really matter. The strongman interpretation is that we just read and skim too much each day.

If so, the title of the article is meaningless clickbait. From reading the article it seems to me the author takes the phrase at face value as they imagine skim reading a third of Middlemarch in a day. I'm surprised they describe different modes of reading as a skill future people will have, since afaik it's normal for most people to approach different types of texts at different speeds and paying attention in different ways. I can skim quickly over comments and articles just building an understanding of the context and basic information and conclusions people are using, while I will read a scientific article anywhere between ½ and 3 pages per hour if I'm actually trying to understand a difficult concept.

That statement makes about as much sense as saying that on average, a human being consumes 4 million droplets of water a day. That is, it's off by 2+ orders of magnitude, and a stupid comparison in the first place. There's nothing to strongman here, IMO.

It's clear what the author meant, yeah, but the fact that they blithely throw around big numbers while clearly not understanding what they mean does affect their overall credibility.

There is no way we skim _billions_ of words every day, on average. Remember one word every second is about 30 years for a billion, so even at ten words a second, you are still taking about 3 years.

So even assuming almost all of that is video, I will still have to question it -- at 2gb/hr that is constantly watching video for more than 16 hours.

It might be accurate if you include every tracker on every article page. If that still doesn't get you there, you can pad it out with youtube and netflix.

And it's not at all clear what "consumes" means here.

If I flick through a news feed of some kind, skimming for headlines that might interest me, am I "consuming" every image and video that scrolls past?

I clicked the link but didn't read the article; have I "consumed" it?

Odd and misleading is WaPo’s MO.

Reply on news.ycombinator.com