Microsoft Word Kurzweil, Ray The Singularity Is Near doc

A Computer's Inherent Dualism.
Neuroscientist Anthony Bell of Redwood Neuroscience Institute articulates two 
challenges to our ability to model and simulate the brain with computation. In the first he maintains that 
a computer is an intrinsically dualistic entity, with its physical set-up designed not to interfere with its logical 
set-up, which executes the computation. In empirical investigation, we find that the brain is not a dualistic 
entity. Computer and program may be two, but mind and brain are one. The brain is thus not a machine, 
meaning it is not a finite model (or computer) instantiated physically in such a way that the physical 
instantiation does not interfere with the execution of the model (or program).
18 
This argument is easily dispensed with. The ability to separate in a computer the program from the physical 
instantiation that performs the computation is an advantage, not a limitation. First of all, we do have electronic devices 
with dedicated circuitry in which the "computer and program" are not two, but one. Such devices are not 
programmable but are hardwired for one specific set of algorithms. Note that I am not just referring to computers with 
software (called "firmware") in read-only memory, as may be found in a cell phone or pocket computer. In such a 
system, the electronics and the software may still be considered dualistic even if the program cannot easily be 
modified. 
I am referring instead to systems with dedicated logic that cannot be programmed at all—such as application-
specific integrated circuits (used, for example, for image and signal processing). There is a cost efficiency in 
implementing algorithms in this way, and many electronic consumer products use such circuitry. Programmable 
computers cost more but provide the flexibility of allowing the software to be changed and upgraded. Programmable 
computers can emulate the functionality of any dedicated system, including the algorithms that we are discovering 
(through the efforts to reverse engineer the brain) for neural components, neurons, and brain regions. 

---

There is no validity to calling a system in which the logical algorithm is inherently tied to its physical design "not 
a machine." If its principles of operation can be understood, modeled in mathematical terms, and then instantiated on 
another system (whether that other system is a machine with unchangeable dedicated logic or software on a 
programmable computer), then we can consider it to be a machine and certainly an entity whose capabilities can be re-
created in a machine. As I discussed extensively in chapter 4, there are no barriers to our discovering the brain's 
principles of operation and successfully modeling and simulating them, from its molecular interactions upward. 
Bell refers to a computer's "physical set-up [that is] designed not to interfere with its logical set-up," implying that 
the brain does not have this "limitation." He is correct that our thoughts do help create our brains, and as I pointed out 
earlier we can observe this phenomenon in dynamic brain scans. But we can readily model and simulate both the 
physical and logical aspects of the brain's plasticity in software. The fact that software in a computer is separate from 
its physical instantiation is an architectural advantage in that it allows the same software to be applied to ever-
improving hardware. Computer software, like the brain's changing circuits, can also modify itself, as well as be 
upgraded. 
Computer hardware can likewise be upgraded without requiring a change in software. It is the brain's relatively 
fixed architecture that is severely limited. Although the brain is able to create new connections and neurotransmitter 
patterns, it is restricted to chemical signaling more than one million times slower than electronics, to the limited 
number of interneuronal connections that can fit inside our skulls, and to having no ability to be upgraded, other than 
through the merger with nonbiological intelligence that I've been discussing.