Hacker News Reader

Comment by jerf | Hacker News Reader

While this may superficially sound like a problem if they get too involved in an evolutionary arms race, I actually don't think this will be a problem. It is extremely unlikely that a robot that is powered on flesh will be able to outcompete robots that get their power much more efficiently and rapidly.

No, the real danger is them deciding they don't need us, or worse, that we are actively in their way, (or worse that our neurons can more productively be used to do something else other than "be a human") not that they are going to digest us.

View on HN

Humans have stuck with flesh long after it was strictly necessary in an evolutionary sense. So I wouldn't be at all comfortable assuming we needn't ever worry about robot taste over-riding necessity.

Personally, I've yet to hear a convincing argument for why machine intelligence would naturally act in a way consistent with its underlying machinery in the first place. But even if we were to assume it would naturally act in such a way, I've also heard no convincing argument for why it wouldn't choose to act in ways counter to that nature, for entertainment, sport, variety or any reason of personal 'taste'. Further, I've yet to hear a convincing argument for why a group of such intelligences would all come to the same conclusions about such topics, even those as seemingly open-and-shut as "whether we should eat other thinking beings, despite a surfeit of better alternatives".

Even if almost all robots decided against eating humans, humans would still have plenty to be concerned about, with regards to those who dissent. Even the dissenters have merely taken a dispassionate stance toward humans accidentally gathered in their metaphorical nets, in a "humans are to robots as dolphins are to humans" analogy.

Humans have stuck with flesh long after it was strictly necessary in an evolutionary sense.

Sure, it's not a strictly necessity, but have you payed attention to how hard it is to be strictly vegan? Evolution isn't going to pick what's morally right, it's going to pick what works best.

I think you're confusing veganism with vegetarianism. Vegetarians don't eat meat, which is doable. Vegans try not to consume or use any animal products at all, which is near impossible. For example: dairy, honey and chewing gum are out of the question, as well as the use of beeswax, leather and products tested on animals.

Veganism is a recent, Western, movement. Vegetarianism has been around for millennia and is practiced by hundreds of millions around the world. In India alone, 40% of the population is vegetarian.

http://en.wikipedia.org/wiki/Vegetarianism_in_specific_count...

Not really hard imho. At least as long as you are living in a major European city, but I would assume North America/some other places should be similar. Have you actually tried?

Don't develop human-like minds, or prove the stability of a goal system under recursive self-improvement. (If you offer Gandhi a pill that he knows will make him want to kill people, he won't take it, because he doesn't want to want to kill people.) If you throw a bunch of crap together and somehow get a mind from it, it's most likely not going to be anything resembling human (and so you'd be right that not all AIs would reach the same conclusions). I'm not an AI researcher so I'll stop here, but you're assuming that many AGIs are going to even have a concept of "taste" or "dissent" let alone a preference for it.

Actually, that's why I carefully stuck to phrasing it in evolutionary terms. I have no idea what AIs may or may not arise, but if the situation has spun this far out of human control regardless of how there may very well be a "species" of robot that desires to eat humans, it won't last long against its robotic competition.

Plus the other robotic competition probably eliminated the humans, anyhow.

Your last point is very creepy, and something I've never considered before. I hope I never become a computer's computer.

Reply on news.ycombinator.com