The brain's 5-million core, 9 Hz computer

"Unlike transistors, neurons are intrinsically rhythmic to various degrees due to their ion channel complements that govern firing and refractory/recovery times. So external "clocking" is not always needed to make them run."<p>Transistors don't need a clock in order to run. They can in fact be set up to create their own clocks. The purpose of clocks is for synchronization across the chip so that we mere mortals can modularize the operation of a CPU. That is, clocks exist mostly so that we can think in terms of sequential gate operations (or from the programmer point of view, assembly code).<p>The author seems to confuse the chosen approach to designing computers (VLSI), and the actual physical capabilities of a transistor. We have opted over the last forty years to develop the CMOS logic gate way of organizing computers. There are other ways, as the brain demonstrates, but it is not clear at all that you can't do it with novel transistor topologies.

Just a small nitpick. While the state of the art in x86 land might be 3GHz, the IBM Power chips and I think the Z mainframes too, not sure on that, have gone far beyond that speed.<p>POWER6 chips reached 5.0GHz in 2008: <a href="http://en.wikipedia.org/wiki/POWER6" rel="nofollow">http://en.wikipedia.org/wiki/POWER6</a><p>POWER7 chips however have been clocked down to 4.25GHz: <a href="http://en.wikipedia.org/wiki/POWER7" rel="nofollow">http://en.wikipedia.org/wiki/POWER7</a>

Hmm, so, if their metaphor really held, the brain's computation could be simulated with a 45MHz CPU? Well, let's fix this up a bit...<p>(1) Give it 1000 clock cycles of cpu work to simulate a single neuro-tick.<p>(2) The clock is actually variable 5-500Hz (from the article).<p>So, 500Hz*5m = 2.5GHz, at 1000 cycles required, so 2500 Ghz of CPU power. An amazon large-instance cluster box is 8 cores of xeon at 2.93 Ghz, about 110 cluster instances to simulate a brain?

This reminds me of Chuck Moore's 144-core, unclocked (!) colorForth CPU: <a href="http://greenarraychips.com/home/documents/greg/GA144.htm" rel="nofollow">http://greenarraychips.com/home/documents/greg/GA144.htm</a><p>I wonder when he'll catch up to 5 million.

At the time of this comment this article has 35 points and 0 comments. Over the past 6 months or so I've noticed a trend that the best HN worthy articles often have points : comments ratios of 5 to 1 or less. This is a clear indication of that.

I've been fascinated by Buzsaki and others in complementary fields since learning of them, and I've also written down some of my own ideas on the subject, from a purely theoretical perspective at least. I can't wait for what we might see in the next few decades.<p><a href="http://guscost.com/2011/04/12/science-analog-confabulation/" rel="nofollow">http://guscost.com/2011/04/12/science-analog-confabulation/</a>

My layman thinking on related subject of building brain-like computer is that currently it's possible (although expensive) to build required hardware - i.e. custom system that will have required number of connected electronic neurons, but I don't see a way to "boot it up". Human brains essentially boots from the DNA - i.e. it's layout and basic functions are predetermined by the DNA and then it gets trained for specifics. Even if we can train machine brain, how do we boot it up to the trainable state?

When we finally design a simulated brain, I wonder if (1) it will be really good at spatial and behavioral behavior (balancing, motor skills, quick non-explicit decision making etc), but really bad at doing math?<p>Not to say we shouldn't keep trying, but we all seem to think that the best computer will evolve from solving lots and lots of partial differential equations into being like the brain, but animal brains have evolved to solve really different questions than why we have been making computers.

Some of the comments seem to revolve around the confusion of maps with territory - CF Lewis Carroll's map "the scale of a mile to the mile" - <a href="http://en.wikipedia.org/wiki/Map%E2%80%93territory_relation#.22The_map_is_not_the_territory.22" rel="nofollow">http://en.wikipedia.org/wiki/Map%E2%80%93territory_relation#...</a>

The brain is also water cooled. Without proper water cooling it overheats causing segfaults and a white screen of death.

transisotr technology cannot emulate brain activity. something new and more analog oriented suits better imho

nature's had a long time to sort out some decent hardware and software configurations. let's follow with computers.

While it's true that some cells, e.g. pyramidals in the hippocampus can exhibit intrinsic oscillations, it's not true for most of the brain. Plus rhythms usually arise in networks, not single cells, and require the network to sustain themselves (that's why for example theta doesn't persist in vitro). Is there an example of a cell that can generate rhythms?