I found the following statement on Richard Dawkins book The Selfish Gene (1989):

"... there are some ten thousand million neurones in the human brain: you could pack only a few hundred transistors into a skull."

Is this statement still true today? Thanks.

  • 2
    \$\begingroup\$ In the Wikipedia article on neurons one estimate is 100 billion neurons. (One idiot wrote that another estimate is 86 billion, as if 100 billion and 86 billion wouldn't be exactly the same :-)) \$\endgroup\$ – stevenvh Jun 13 '11 at 8:29
  • \$\begingroup\$ Someone asked in a comment to Leon Heller's answer how many other transistors a typical transistor is connected to in a microcontroller, but he deleted that comment. I find it an interesting question. Does anybody have an idea? \$\endgroup\$ – Federico Russo Jun 14 '11 at 14:28

It wasn't even true back then. Well, maybe that's why Dawkins is a biologist and not an engineer. :-)
Today's processors pack billions of transistors on a die a few square cm in area and less than a mm high. There would fit hundreds of them in a skull, maybe \$10^{12}\$ transistors.
Even if you look at discrete transistors there would fit more than just a few hundred. I guess SOT-23 already existed in 1989, and then you would get \$10^5\$-\$10^6\$ of them in a skull.

edit (2011-06-13)
I own a copy of The Selfish Gene, and was curious what Dawkins had in mind, so I looked into it. Her's more from that paragraph:

The basic unit of biological computers, the nerve cell or neurone, is really nothing like a transistor in its internal workings. Certainly the code in which neurones communicate with each other seems to be a little bit like the pulse codes of digital computers, but the individual neurone is a much more sophisticated data-processing unit than the transistor. Instead of just three connections with other components (sic), a single neurone may have tens of thousands. The neurone is slower than the transistor, but it has gone much further in the direction of miniaturization, a trend which has dominated the electronics industry over the past two decades. (The Selfish Gene, p.49)

Somebody must have told Dawkins that a transistor has 3 pins :-).
Anyway, he doesn't only compare the numbers of neurons (or neurones, BE?) to transistors, but also points out that the neuron is a lot more complex, partly because of its thousands of connections. My guesstimate is that you'd need \$10^5\$ to \$10^6\$ transistors to emulate one such neuron (maybe as an analog instead digital computer?). Which means that a skull stuffed with GPUs wouldn't still come close to the processing power of a brain.
And then there's the problem of all these connections. They're the real power, not just the large number of neurons. We don't have the technology to build such complex systems, and IMO won't for a long time. And then I'm not even talking about the dynamic nature of these connections: they can rearrange themselves, making new connections and breaking others.
To put all these AI suckers in perspective, take a look at our vision system. In a second we can process a stereoscopic image of \$10^8\$ pixels, create a virtual 3D model of the scene and identify objects in detail. Move half a meter to the right and you add lots of new data. There's still a long way to go...

  • \$\begingroup\$ Moore's law (double every 18 months) tells me it must have been on the order of 10^6 per die then, so 10^9 in total, maybe 10^10. Though cooling would of course be a major problem. \$\endgroup\$ – starblue Jun 11 '11 at 20:21
  • \$\begingroup\$ It seems like you could accomplish the output function of a neuron, which appears to be a sort of logarithmically saturating amplifier, with no more transistors than a regular op-amp - but the input functions are, I think, where the transistor count would soar, (and neurons have LOTS of inputs) owing to the fact that, at least as far as we understand them, they seem to be variable gain units each with their own independent integrator ('weight'), so probably another op-amp or two's worth of transistors right there. Times 10^3-10^4. And that's a gross simplification. \$\endgroup\$ – JustJeff Jun 13 '11 at 11:32
  • \$\begingroup\$ I gained an extreme appreciation for human intelligence when I was working on a bipedal walking algorithm last semester. Something we take for granted in even the least intelligent humans is absurdly complicated to code/model, let alone actually develop. \$\endgroup\$ – NickHalden Jun 16 '11 at 19:27

Where neurons score over transistors and electronic devices is the vast number of connections they make to other neurons - 7,000 on average.

  • 1
    \$\begingroup\$ yeah, it'd not so much be the devices you'd have to concern yourself with, it'd be all that wire. \$\endgroup\$ – JustJeff Jun 11 '11 at 20:47
  • 1
    \$\begingroup\$ Incedentally, modern FPGAs usually have the same problems: They're not gate-limited, they're wiring-limited. \$\endgroup\$ – Kevin Vermeer Jun 13 '11 at 2:09

The 68K coincidentally had 68,000 transistors in it, and that chip was available in 1979. You could certainly fit several 68K dies in the same space as a brain, and thereby exceed "several hundred" by three orders of magnitude, with what would have been 10 year old technology at the time of the statement. Perhaps if you went with TO-92 packages, you might not quite get to 1,000 of them.

OTOH it should be pointed out that a decent model of a single neuron would probably involve more than a single transistor.

  • 1
    \$\begingroup\$ No doubt this (the TO-92 package) is the "transistor" the author was thinking of. \$\endgroup\$ – dmckee --- ex-moderator kitten Jun 12 '11 at 0:50
  • 2
    \$\begingroup\$ Brain volume: 1400 ml (according to "Principles of Neurosurgery", Elsevier Mosby, 2005). 1000 TO-92, neatly stacked in X, Y, Z directions (so less than optimal): 420 \$cm^3\$. That's 3300 of them in a brain volume. \$\endgroup\$ – stevenvh Jun 13 '11 at 7:18
  • 1
    \$\begingroup\$ @stevenvh - you know, that's not entirely surprising. I didn't bother to look up the numbers precisely b/c I suspected that Dawkins was probably exaggerating for effect. Apparently, we are left to surmise that the only transistors Dawkins knew of were TO-3's =P \$\endgroup\$ – JustJeff Jun 13 '11 at 11:24
  • 1
    \$\begingroup\$ "surmise that the only transistors Dawkins knew of were TO-3's" - I'm not even gonna touch that :-) \$\endgroup\$ – stevenvh Jun 13 '11 at 11:39
  • \$\begingroup\$ TO-3, like the 2N3055! hFE of 5 (at 10A) \$\endgroup\$ – Federico Russo Jun 13 '11 at 13:27

This is a misquotation! The Selfish Gene was not written in 1989. It was written in 1976!

What Richard Dawkins published in 1989 was the second edition of the book. Indeed, this second edition includes endnotes where he updated the data about transistors:

"...my remarks about [computers] have become [...] dated. [...] The number of transistor-equivalents that you could pack into a skull today must be up in the billions.".

Dawkins made his homework before writing about transistors, you didn't make yours before quoting him...

  • 2
    \$\begingroup\$ becko didn't claim that the 1989 he refers to was the first edition. And it's not a misquotation; it's exactly what the book says. If RD wants to be sure that the reader doesn't miss the endnotes, he should add references to them. \$\endgroup\$ – stevenvh Jul 8 '11 at 2:36
  • \$\begingroup\$ @Alonso: I missed that endnote. Thanks for bringing it to my attention. \$\endgroup\$ – becko Jul 8 '11 at 4:10

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.