More Fun With Numbers – The Information Overload Edition

info overload
An article in yesterday’s NYT says we’re producing digital information in volumes that will soon exceed our capacity to store it. Something about these huge numbers didn’t ring right. “Where is the knowledge we have lost in information?”, TS Eliot once wrote. Do we have the capacity to actually use all this digital exhaust? Is it all worth producing, and is the ease of producing and storing it just making it harder to find the stuff that’s actually useful? And for all the zeroes after these numbers, is this really a lot of information for nearly seven billion humans to be producing? I decided to do some math to find out.

Since we’re speaking in large round numbers, there are about 5 x 1030 bacteria on Earth. Even if these remarkably complex creatures only produced one byte of information each in their lifetimes, their total information production would be 30 billion times the aggregate human output of a mere 161 exabytes. So by comparison with bacteria, we humans are still junior league information producers.

Our bodies are also pretty good information processors compared to computers. A recent study claimed that our bodies process 2MB of information per second (most of it unconscious or subconscious) or 5 x 1015 bytes of information in a lifetime. So all human bodies currently on the planet process 5 x 1023 bytes of non-digital information each year, or about 2500 times as much as the digital information which we are collectively producing and which all our machines are collectively processing, storing and distributing. But then our machines are pretty dumb and slow compared to the marvel of the human body.

Even more remarkable, the conscious human mind is only able to absorb an average of 3 bytes of information per second over a lifetime, or 7GB of information in an entire lifetime. That means the 6.7 billion humans brains on the planet are only able to absorb 6 x 1017 bytes of information in a year, of which at least 95% is sensory or interpersonal (i.e. non-digital), so collectively we are absorbing only 3 x 1016 bytes of digital information in a year, increasing by our population growth rate of about 2%/year. Meanwhile the amount of digital information we are producing (presumably in the hope others will somehow use it) is currently (according to the NYT article) 161 exabytes or 2 x 1020 bytes of information per year, growing at over 50% per year. So we are already producing 6000 times as much digital information as we’re consuming (1500 times as much if you exclude duplicates/copies of information), and by 2010 we’ll be producing 30,000 times as much digital information as we consume (7000 times as much if you exclude duplicates/copies).

That means at least 1499/1500 (99.93%) of the digital information being produced by us and our machines will never be read or consumed or otherwise used by any human. We are producing and capturing it “just in case”. And an increasing amount of the digital information we produce is designed to be read andused only by other machines.

Just as well I guess. My head’s already full.

Thanks to my colleague Gabrielle Gaedecke for the link. Graphic from

This entry was posted in Using Weblogs and Technology. Bookmark the permalink.

3 Responses to More Fun With Numbers – The Information Overload Edition

  1. dataguy says:

    “…our bodies process 2MB of information per second” yet “…the conscious human mind is only able to absorb an average of 3 bytes of information per second”. May be that data is correct because that ratio makes no sense to me. Even more confusing to me is how they could make these measurements since our bodys are not binary devices. I have no background in this area but I do not believe that 3 bytes/second can possibly be correct.

  2. Ken Whitley says:

    Indeed – 3 bytes/second seems ridiculously low. If you compress,say, a piece of music until it starts drastically losing quality, you’re still hearing tens of thousands of bytes worth of audio data per second. Granted, you aren’t extracting every possible bit of data from the music, but you’re hearing the music, and the lyrics, and noticing the nature and quality of the distortion and artifacts. And you may be managing ten or twelve chat conversations at once on your computer, while talking aloud to your partner, noticing what you feel as you stretch your neck and back, solving some business problem in your head, and smelling that the back door is open, the brownies in the oven are nearly done, and your dog is sneaking up behind you.Myself, I’ve been a speed-reader all my life, and from personal experience I basically agree with the theory of speed-reading I came across later in life, that suggests we can take in about seven CHUNKS of information per second, and the secret to reading faster with full comprehension is to take in LARGER chunks at once, rather than trying to take them in faster.Anyone who works professionally with images can compare the size of file needed to accurately contain images without compression with the amount of information a professional can take in from an image in a single glance – likewise anyone who spends significant time outdoors and evaluates their environment as a whole. Sure, you can fool people with compression, basically by getting them to make *assumptions* about what they see. But getting them to settle for less data is not the same as not being able to take it in.The three bytes/second MAY be accurate if the definition is “how much random meaningless data, presented in a sensory-unfriendly format, can a person perfectly memorize?” i.e. strings of random digits.Another counterexample – compare the “compare two pictures and find the differences” puzzles with the real-life experience of walking into a personallyimportant room and noticing instantly that something is awry. How much data does it take to scan a whole room and notice the scissors aren’t where you left them?

Comments are closed.