People have long cared deeply about what happens to the characters in epics, dramas, stories, films, television--and now video games. Art produces artificial characters (that's what art does), and the more they seem believable and interesting, the more we are likely to care about their artificial lives. We may care a lot about an artificial character if we have "cathexis" with that character. Cathexis = the concentration of mental energy (to a possibly unhealthy degree) on one particular person, idea, or object. We might then impute to the artificial life a definite reality. The NPC can seem to become "real". Just like Frodo did. Or Harry Potter. Or Darth Vader.

One wouldn't want to get carried away with imputing reality to characters in games and epics. People get carried away with imputing complex mental processes to their pets. At least an actual dog can have actual preferences (our retriever had all sorts of preferences and balked when they were not honored). A character in a novel can not have actual anything, EXCEPT that the author is an actual person. The same goes for video games -- there is a real author behind the artificial characters.

Part of the pleasure of fiction is to join with the author in a game of "let's pretend"--the necessary voluntary suspension of disbelief. Of course Frodo, Harry, and Darth don't exist. They are nothing more than ink spots on paper. BUT... we can "bring them to life" and sort of "share their reality". We take the author's text as a testament of their reality.

Preference is psychological and/or biological. The letter "A" and NPC's are neither of these. The latter is a constructed artificial imitation.

The argument presented by the essayist is that NPCs can display actions to try to prevent "death" (i.e. a sleep state), which is argued to be not-preferable to the NPC.

Aren't we just machines as well?

I'd say we are most certainly not machines, nor is our mind a super-advanced computer. This isn't a difference of degree, I think, but of kind. So there is no point where a program becomes a preference, because preferences are not built up of programs.

I think we differ completely, then, on our conception of the mind. I come from a computationalist perspective.

Why choose that perspective though?

I think the choice involved in having a preference of one thing over another is not the same choice as a program can make. Preference suggests a transitive range of feelings. You prefer to drink tea over coffee and to drink coffee over milk. A program can be made with filters to mimic the transitivity of these preferences, but its filters unlike my preferences don't change. I can choose to drink milk today instead of tea, because I feel like it, regardless of my stated preferences.

And, given the ordinary meaning of the word "preference" and its connotations, that's a category error. Likewise in calling us machines.

From a computational perspective, humans are just like computers. But it's immoral in many cases to stop someone from pursuing their preferences. — darthbarracuda

These have the appearance of substantial claims but both fall flat. The first claim falls apart by recognition that it is a likeness by analogy. Humans are not just like computers, but we can make sense of cognition by making an analogy to computers. Analogies are only as strong as they approach equivalence, but in this case, the difference between humans and computers is precisely what we are after so we cannot use the analogy to make sense of the difference in cognition.

The second claim has a bit of a weasel word that intimates a strong position, but actually fails to take any position at all. By including the word "many" you implicitly suggest there is something inherent about the impermissibility of preventing someone from pursuing preferences, but "many" can also be used to imply that there is at least one case where it is not immoral to stop someone from pursuing preferences. We need to know what the salient features are that make this immoral or not before we can make a judgement about NPCs.

From a computational perspective, humans are just like computers. — darthbarracuda

Take some random object in your hand, throw it up in the air, and then catch it. A computer, I understand, has to do some pretty fancy math to get a robotic arm to do that. But you didn't do any math when you caught it. We may do the same things that computers do, but the stuff going on in our brains is not like the stuff going on in the computer in some important ways.

In some ways the computer-brain analogy may be extended to incoherency.

But it's worth noting that the parallelism of many modern day computers is commonly seen to be that of the brain. Parallel computers can compute things that a serial computer would struggle to do so.