The Black Vault

at1with0 wrote:Self-modifying code is self-aware already. It needs to be in order to be able to change itself.

Machines will achieve human-level intelligence by 2028 (median estimate: 10% chance), by 2050 (median estimate: 50% chance), or by 2150 (median estimate: 90% chance), according to an informal poll at the Future of Humanity Institute (FHI) Winter Intelligence conference on machine intelligence in January.

at1with0 wrote:Self-modifying entails self-awareness.
It has to be aware of what it is in order to change what it is.
Did I ever say that self-modifying code has human-like sentience?
Sentience and self-awareness are not identical.

Most experts would not consider that a robot is really selfaware
just because it can visually recognize its own motion
or itself in a mirror since a program specifically designed to
achieve that kind of recognition without having a genuine
awareness capacity can be developed. The ability to visually
recognize one-self is not enough for achieving selfawareness.
Self-recognition can be a side-effect of selfawareness,
but not a pre-requisite. We believe a robot needs
a capability to attend to its internal states in order to be selfaware.
Current approaches as described in the previous
section do not focus on directing a robot’s attention to its
own internal processes. If we add an attention process to a
robot so that it can focus on processes that happen internally
during self-recognition activities then we would consider it
to be self-aware. What is crucially important is not the
ability to recognize itself in a mirror (e.g. a visual inverted
reflection), but rather to be aware of its own emotions,
perceptions, beliefs and intentions during the recognition
process. If a robot has totally lost all of its outward facing
sensations, it may not be aware of its environment (external
awareness), however it can still be aware of itself (selfawareness).