Recursive Self-Improvement

From Lesswrongwiki
Revision as of 02:02, 15 June 2012 by Daniel Trenor (talk | contribs) (See Also)
Jump to: navigation, search

Recursive Self-Improvement is an approach to Artificial Intelligence that allows a system to make adjustments to its own functionality resulting in improved performance. The system can then feedback on itself with each cycle reaching ever higher levels of intelligence.

Assuming a machine of sufficient power was available, once an initial “escape velocity” had been reached the intelligence of such a system would be expected to accelerate rapidly and according to Eliezer Yudkowsky become capable in a matter of “weeks or hours” of achieving what would take humanity decades.

Ultimately such a system would be bound by the limitations of it's hardware. The software would be searching for cognitive algorithms that were the most optimal.

Blog Posts

External Links

See Also