>Penrose, for example, thinks the mystery may yield to a deeper understanding of quantum mechanics.
OK.
> Higher level thinking is mostly a strategy module for the "next 15 seconds" control system.
I'm not convinced we have even that much of a clue about consciousness.
isn't this guy's bias. he's, in part, interested in how the AI works, because he's interested in understanding how people think.
i'm also totally sick of the buzzword.
to reveal my bias, here's another thing AI doesn't do but we do all the time: have feelings.
Check out various algorithms to do so, such as [1]. Model building from sensory input is very simple, and in my opinion only tangentially related to consciousness.
For humans, the model is bound externally in parental memory, media influences, and impersonal record systems.
But it's very much there as external programming. Baby humans don't very well when it's missing.
This is one of the basic activities AI researchers have been solving for years.
In many cases, this is not even a hard problem.
"We share a conviction that no philosopher or scientist living today has the foggiest notion of how consciousness, and its inseparable companion free will, emerge, as they surely do, from a material brain."
That's reasonable enough. Nobody seems to have a clue about how "consciousness", whatever it is, really works. If someone did, it would have been implemented in a computer by now.
"We mysterians are persuaded that no computer of the sort we know how to build—that is, one made with wires and switches—will ever cross a threshold to become aware of what it is doing."
Nah. Now he's just mouthing off. There's a long history of this sort of thing from philosophers. Aristotle claimed humans were special because they could do arithmetic. Now we know that arithmetic is simple, in the sense that it takes a very small parts count. Hubert Dreyfus, in 1972, said computers couldn't play even a decent amateur game of chess. Now, any laptop PC can trounce grand masters.
What AI isn't doing well yet is predicting the consequences of actions: "what will happen next", and "what will happen next if this decision is made". Basically, the technology to get through the next 15 seconds of life. This is the beginning of common sense, and really ought to be worked on more. We have the tools now to go at that problem. Self-driving cars have to do some of this, and answers may come out of that technology. This is an essential technology for robots which operate in an unstructured world.
Once that problem has been solved, we might be in a position to talk about "consciousness" from a non-clueless perspective. Higher level thinking is mostly a strategy module for the "next 15 seconds" control system.
At least the article doesn't use the world "quantum".