KurzweilAI.net reports the following: Nanochip Emulates Human Brain
Researchers are developing a new 3-D chip design using spintronics and a complex interconnected network of nanowires, with computing functions and decisions performed at the nodes where they meet — an approach similar to neurons and axons in the brain.
It combines the storage capability of a hard drive with the low cost of memory cards, potentially increasing memory capacity by 200 times from an average of 500MB to around 100GB.
In my opinion, this is a pretty big deal. It is also a very good illustration of a certain scenario that I have written about in my Singularity FAQ.
In the FAQ, I describe the enabling technologies for a fullblown superior artificial intelligence. Two very important ‘enablers’ are nano-circuitry, and a thorough understanding of how the human brain’s intelligence works.
A nanochip that emulates the human brain to store information efficiently, is an impressive feat. It not only shows that nanocircuits are possible (as if there was any doubt in these last few years), but it also (once again) shows that it is entirely possible to understand how our own brain works.
It is an impressive achievement. And yet this is only the beginning. We ain’t seen nothin’ yet. I expect that chips will slowly but surely take on more and more characterstics of brains that are based on the neural network paradigm. Such a chip could dynamically (re)configure itself to be faster at specific tasks. Standard CPU’s, the ones we use today, can’t do that. They need to be very general in design… being able to run any sort of application, to reflect the demands of users worldwide.
This makes them slower than, say, CPU’s in game consoles, which are dedicated to run graphical applications. Games that run on game consoles require CPU’s that have much less clockspeed than games that run on general-use PC’s, due to the game-dedicated hardware of game consoles.
Nano-tsunami also reports on this: New Microchip Design Could Be The Key To Expanding Mobile Phone Memory.
Researchers from Imperial College London, Durham University and the University of Sheffield say their new computer chip design will enable large amounts of data to be stored in small volumes by using a complex interconnected network of nanowires, with computing functions and decisions performed at the nodes where they meet a similar approach to neurons and axons in the brain.
Large amounts of data in small volumes?
Sounds like a technology that could help out nicely with The Future Of Computers.
Another fine example of a respectable amount of data (4GB) in an extremely small volume, is the new iPod Nano. The Archos GMini 402 is not too shabby either: it has 20GB of space, it can play music, movies, videogames, and it functions as an external harddisk.
I’d say a true ‘all-in-one’ pocket-machine is just a few years away.
Back on topic. By reading the title of the article, you’d think that this is all about memory for mobile devices. But that’s just one of many optional applications:
Lead researcher Russell Cowburn , Professor of Nanotechnology in Imperial’s Department of Physics, explains: “The new video mobile phones are very popular, but they desperately need more memory so that people can take longer videos and store them. This technology has the potential to transform mobiles into fully functioning video cameras, in addition to a range of other applications.”
I’d say we’re headed for a very cool mobile and wireless world, where we take our computational power with us wherever we go.
Extrapolations show that, circa 2015/2020, we’ll be able to record our entire lives at only a fraction of the storage space that will be cheaply available to us.
Think about that.