Hacker News Reader

Comment by munificent | Hacker News Reader

> Does the in-memory simulation contains a model of the in-memory simulation?

Well, yes. I'm not sure about Pong specifically, but in many games with computer controlled entities, the game AI for each entity simulates what the entity "knows" about the world, which is different from the simulated world state itself.

(Also, at least according to Wikipedia, sentience is distinct from self-awareness. The former only requires the existence of a subjective experience, not necessarily self-reflection.)

View on HN

I also thought about AI in games, but for it to be equivalent, the AI should be able to have perception and/or control over the state of the AI itself, not merely about other game entities. E.g. some check like "stop reasoning if the inference process is taking too long to find an answer" is the kind of loop that would a process to have some kind of awareness about itself (specially if this was emergent behaviour, and not just a low-level pre-programmed test).

I'm not sure I buy the "sentience without self-awareness" concept. First, defining it as "subjective experience" seems circular, and second, reflex actions like e.g. plants moving towards the light or reacting against a parasite would count as subjective experience; in the extreme, even a blind chemical reaction like fire could fall under that definition.

> I'm not sure I buy the "sentience without self-awareness" concept.

Honest question -- what do you think is going on inside of babies and children? Self-awareness as a developmental process is complex and relational, and takes many years. Our senses though are more or less put together by puberty, built up by simple interactions with the world.

Rational self-awareness, sure. But I believe higher animals and babies have self-awareness in that they are conscious of their own perceptions. It's difficult to put it in words without falling in circular definitions, but the key feature would be higher brain features working on top of the low-level automatic reflex information processing.

Simpler animals like wasps and worms don't seem to have those, or have really simple versions; their behaviours look more like following hardwired programs reacting to specific conditions, not influenced by some inner "mind state".

Reply on news.ycombinator.com