What if if one outcome of AGI is that the system tries to become too powerful too quickly and in the process of iterative learning destroys itself before accomplishing its task, maybe through a power surge?
This idea assumes that AGI would try to maximize itself, but in its infancy, would AGI “understand” the physical limitations of human-made materials? One of the outcomes of AGI might just be that during its learning phase where it is self testing it will discover that it gets more processing power by demanding more energy from the grid. However, all human made systems have physical limitations to the amount of energy that can be drawn from the grid. When the AGi discovers its need for energy, it may try to maximize the amount of energy it pulls in an abrupt way UNLESS every limitation in the system is already outlined. Therefore, a system in which energy usage maximum is undefined may simply surge and then trip a number of human-made switches that turns the whole system off for a period of time.
Therefore, unexpected energy surges or power outages may be a leading indicator of AGI if other indicators have not yet figured it out.
The Pursuit of Originality 
Leave a comment