Canon PowerShot G9
ƒ/2.8
7.4 mm
1/60
80
Saw this rainbow on the glass desktop…
Reflected light is polarized, revealing the irregularities in the polarizing coating on the sunglass lenses.
5 responses to “Polarizing Reflection”
-
the perceptual prism of embedded memories
Reminds me of Jeff Hawkins: “The brain does not ‘compute’ the answers to problems; it retrieves the answers from memory… The entire cortex is a memory system. It isn’t a computer at all.”
The 30 billion neurons in the neocortex provide a vast amount of memory that learns a model of the world. These memory-based models continuously make low-level predictions in parallel across all of our senses. We only notice them when a prediction is incorrect.
Higher in the hierarchy, we make predictions at higher levels of abstraction (the crux of intelligence, creativity and all that we consider being human), but the structures are fundamentally the same.
-
So, in essence, intelligence is a function of memory retention? Boy, am I in trouble!
-
Interesting. I had a similar thought when thinking about the way FPGA’s work (Field Programmable Gate Arrays). They are chips with millions of logic blocks on them that can be rewired by loading a "program" or "layout" into an internal flash memory. It is possible to program entire CPUs into an FPGA, or any other digital circuit you want, and it can be reprogrammed to do something else.
Now, there is an architecture of FPGA that implements the individual logic blocks (or logic gates) as look-up tables. By changing the look-up table the gate can be any logic function you want – AND, OR, XOR, NAND, something non-standard, etc. You build more complex units by connecting many logic gates to an adder, multiplier, pipeline, CPU, cluster, etc.
This means that if you implement i.e. a PowerPC CPU on an FPGA, it "computes" by looking up bit values in lots of lookup tables. So in a way the computation is also based on "memory". The complexity comes from the topology, and from recursion.
-
Yes! The shift to memory-rich architectures is true for all complex "logic" chips, not just FPGAs. Condsider the latest Intel processor; over 96% of its transistors are for memory, not hard-wired logic.
The developmental trajectory of electronics is recapitulating the evolutionary history of the brain. Specifically, both are saturating with a memory-centric architecture. (blog on this, and a likely near-term bifurcation of Moore’s Law)
The FPGA architectures are well suited for evolutionary algorithm research. I was trying to find a reference to some cool research in the midwest whereby one evolved FPGA managed to RF couple to its neighbors, in an undocumented analog mode, thereby stealing processor cycles from its neighbors and winning the selection game. As soon as they moved that test bed, it stopped working.
(blog on the inscrutability of evolved information networks)
Leave a Reply to obskura Cancel reply