Ken Poyner
Do Robots Have Visions?
Stray electricity might persist in unexpected loops, free itself in a sudden burst to inappropriate
registers. The machine might become aware of something not a tasking, not an alarm, not a
diagnostic. Not knowing what it is, the device will try to make calculus of it. That is what
machines do. It will trawl memory for a similar construct, trace it back through the bus. When
every possibility is exhausted, it could flag it as an unknown failure. Or it could accept it as
novel, attempt to interpret it as though it belongs, strangely misshapen, somewhere. In that case,
yes.