Cell Imaging
After reading Kurzweil's section on brain scanning and how it is exponentially-accelerating, it dawned on me that I am actually pretty-well suited to refute Kurzweil on these matters. As it so happens, I just wrote a proposal to a certain government agency on a new method for imaging plant cells.
So the first problem with Kurzweil's claims is that all optical techniques, including the knife-edge scanning microscope he cites, are limited by the diffraction limit of light:
Spot Size = 0.61 lambda/N.A.
where lambda is the wavelength of light and N.A. is the numerical aperture of the microscope objective. N.A. will be around 1 (up to 1.5 for immersion lenses) and the wavelength is usually no lower than about 200 nm. So the minimum spot size will be around 200 nm. This resolution will allow you to see some cell structures but it won't give you the resolution to see the interneuronal processes which are important to cognition. (Much of Kurzweil's religion is based on the belief that you can ignore what happens within neurons. I will debunk this faith-based assertion in future posts.)
Techniques based on AFM (atomic force microscopy) require that the sample be on a flat substrate and can only measure surface structures.
Actually, the main problem for in vivo brain scanning is that neurons move tens of microns whenever the heart beats. All imaging techniques require relatively stationary samples so all these techniques will fail in vivo.
So, of course, Kurzweil falls back on his belief in the inevitability of nanobots, even though there isn't a single scientific article on nanobots. Not ONE!
So, once you scratch the surface, all of Kurzweil's arguments fall apart.
So the first problem with Kurzweil's claims is that all optical techniques, including the knife-edge scanning microscope he cites, are limited by the diffraction limit of light:
Spot Size = 0.61 lambda/N.A.
where lambda is the wavelength of light and N.A. is the numerical aperture of the microscope objective. N.A. will be around 1 (up to 1.5 for immersion lenses) and the wavelength is usually no lower than about 200 nm. So the minimum spot size will be around 200 nm. This resolution will allow you to see some cell structures but it won't give you the resolution to see the interneuronal processes which are important to cognition. (Much of Kurzweil's religion is based on the belief that you can ignore what happens within neurons. I will debunk this faith-based assertion in future posts.)
Techniques based on AFM (atomic force microscopy) require that the sample be on a flat substrate and can only measure surface structures.
Actually, the main problem for in vivo brain scanning is that neurons move tens of microns whenever the heart beats. All imaging techniques require relatively stationary samples so all these techniques will fail in vivo.
So, of course, Kurzweil falls back on his belief in the inevitability of nanobots, even though there isn't a single scientific article on nanobots. Not ONE!
So, once you scratch the surface, all of Kurzweil's arguments fall apart.
Nanobots not required.
Imaging of live brain function at the protein level
http://advancednano.blogspot.com/2006/07/singularityai-related-responses-of.html
Scanning optical areas at 10-30 nm resolution over large areas
http://advancednano.blogspot.com/2007/08/optical-scanning-larger-areas-at-10-30.html
Also, see metamaterials and superlenses for optical viewing at less than diffraction limit.
Reply to this