[ExI] Critical take on The Age of Em

Rafal Smigrodzki rafal.smigrodzki at gmail.com
Thu Jul 21 17:05:20 UTC 2016

On Thu, Jun 23, 2016 at 5:33 AM, Robin D Hanson <rhanson at gmu.edu> wrote:

Yes many kinds of software today don’t do much learning, but we do have
> many real systems that do learn, usually based on statistics. (We also know
> of many software systems where the object code is much larger than the
> source code for other reasons.) So if you think that makes a huge
> difference, then imagine instead that someone gave you dozens of examples
> of statistical analysis code from decades ago. You’d have the same problem.

### Again, this is reasoning by analogy. I can only again caution against
trusting conclusions based on this kind of reasoning.

> You don’t know that human brains are any more modular than is typical
> software. You don’t know that it only embodies a small number of
> principles, without masses of other implementation details also required
> for it work. And you don’t know that the many different parts of the brain
> are all written in the same “language”.

### To the contrary. There are already many BMIs that meaningfully access
both input and output from very diverse cortical areas relatively high in
the hierarchy (visual, motor). This implies that the cortical language is
accessible, it is accessible throughout the cortex, and it embodies a
sufficiently small number of principles as to be accessed even by our still
crude technology.

This one datum doesn’t indicate that the brain is any more modular that
> most software systems. Saying it is a “different type” and thus vastly
> easier to understand and modify just sounds like wishful thinking.

### Saying "different type" is a conclusion, not a premise of my reasoning.

Here is another neat datum: The "movies" of visual imagery reconstructed
from fMRI images of human cortex. This directly implies that human mind
states are highly accessible, and reasoning by analogy is not needed to
reach this conclusion.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.extropy.org/pipermail/extropy-chat/attachments/20160721/eb32c9db/attachment.html>

More information about the extropy-chat mailing list