Created: August 16, 2015 / Updated: May 6, 2018 / Status: in progress / 3 min read (~583 words)
An AGI, like a human being, can be focused to work on pretty much anything. This is both a good and a bad thing, in the sense that you can have it work on very complex topics that will improve our understanding of the world, but at the same time it can be tasked to do terrible things, which are completely opposite to that first objective.
Thus, it is important to determine what are the aspects that make an AGI practical and working toward appropriate goals vs one that is simply executing code with no definite purpose.
The goal of this study is to attempt to discover what are the critical components of an AGI that would make it useful to humans in our everyday life.
- Deconstructs incoming stimuli.
- Matches the stimuli with already perceived stimuli patterns.
- Evaluate existing responses and offer alternate responses if the current ones are unsatisfactory.
- Records the selected response for further analysis.
- Executes the selected response.
- Records the result of the response on the environment (other agents).
- Audio amplitude is obtained.
- High amplitude audio stimuli are queried in the KB.
- Current responses are "doing nothing", which is satisfactory.
- Decision is stored (use count incremented).
- AGI does nothing.
- Environment does not change.
- Internally generates desires.
- Determines goals and objectives to fulfill those desires.
- Plans a path to reach those goals and objectives.
- Decomposes a goal/objective into sub-goals/objectives until concrete actions can be executed.
- Executes the necessary steps to reach those goals and objectives.
- Evaluate if the taken steps have brought it closer of further from its initially defined target.
- Re-adjusts its plan as necessary.