Well, you know, some people lead pretty interesting lives, and maybe it would be nice to experience their experiences or to record everything], not just node]the odd thing. How much space would it take to fit an entire human life on a hard drive?
In 1986, T. K. Landauer did a study that estimated that people only take in and remember about 1 byte per second. At that rate, assuming an 80 year life span, you’d only need 80 x 365.25 x 24 x 60 x 60 bytes or a 2.4 gigabyte harddrive to store your entire life experiences.
That seems a bit low to me, so I thought, what about if you wanted to store everything the nervous system receives, not just what it remembers, or just the stuff that gets to the brain, but everything, so that if you rigged someone up to a replay device they would have exactly the same experiences for their entire life.
By current estimates, there are approximately 380 million receptors in the human body. These are just the neurons that take something from the outside world and turn it into a signal that can go into the nervous system.
Usefully enough, these receptors, although they fire many peaks when excited are only on or off, so the analog to digital conversion is as simple as you can get – only 1 bit is needed per receptor.
The next task was to decide on a sample rate for the life dump recording. The sample rate should be high enough so that two stimuli in quick succession that the body can tell apart would appear as two stimuli in the life dump. The receptors can only fire at a maximum of 1200 Hz. Initially I thought that I would use this figure as the sample rate but then I realised, although this is the maximum firing rate, the receptors can fire at different times – they aren’t clocked synchronously at 1200 Hz. The figure I really need is the smallest time interval that two stimuli can happen together for the biological system to perceive them as two separate events. It’s kind of a resolution in time. Bear in mind the auditory system can distinguish between sounds even near the 10kHz mark.
The figure that I decided to use in the end was the transduction time in visual receptors, taking this as an average across the whole body (seems reasonable, approximately 70% of all receptors are visual ones anyway). The transduction time is the time it takes for the receptor to turn an external signal (heat/light/pressure) into a physiological one, so I figure this is plenty quick enough. It’s probably a lot faster than I need, but it’s the best figure I’ve got so I’m going to use it. The visual transduction time is about 6 femtoseconds or 6×10-15 seconds, which gives us a sample rate of 1/6 femto seconds or 1.7×1011kHz.
Now multiplying those values together gives us the number of bits per second we need.
380×106 (receptors) x 1 (bit per receptor) x 1.71011kHz (sample rate)
which is 7.4 zettabytes per second. That’s about 1470 million times as big as google cache. Too big you say? Well according to one estimate, there will be 1 zettabyte of information on the world wide web by 2010.
Well lets work out the figure for an 80 year life. 7.4 zettabytes s-1 x 80 x 365.25 x 24 x 60 x 60 = 18 700 000 000 zettabytes for a lifedump.
Bearing in mind the kind of data, I think I can safely say we would expect to get some pretty fantastic compression rates on it, but even without compression I reckon we’ll have the kind of storage necessary to make a life dump within 150 years.
Please contact me if you can improve this calculation in any way – I’d particularly appreciate a better figure for temporal resolution in receptors.
First published on everything2.com (hence the linking) under the name delfick.