They certainly don't have a lot of programming experience. I can see them developing something with a bunch of complexity, experiencing a bug, being unable to i

They certainly don't have a lot of programming experience. I can see them developing something with a bunch of complexity, experiencing a bug, being unable to identify what's causing it (partially due to not having good tools), then declaring it a "intrinsic part of the program"... or "feature". Maybe not that far off after all... The article finally lost me with "The only way we can find out about the consciousness of others is by asking them [..]". So just add the appropriate response to a sufficiently capable chat-bot and we are done? And if more is required than just a simple yes/no, the same checklist can be used to build something artificial. But since consciousness is one of the good old "you know it when you see it", we will likely shift the goal post pretty hard as AI development progresses, and mostly use the term to exclude what we achieved so far. Kind of like we stop calling something AI once we achieved the a task and use a more technical term... e.g. the many CNN's in a Tesla. Hope I get to experience much of that in my lifetime.