Hacker News Reader

Comment by gear54rus | Hacker News Reader

To all those saying it is impossible:

The point you are making about it being impossible is as moot as the one those that fear AI are making. From logical point of view, singularity (a moment when AI has modified itself to the point beyond our capacity of control or even understanding) seems possible. I don't see a way for us to tell what will happen when it hits.

One needs to remember that AI itself may not be evil, it just may exterminate the human race because it feels it is the right course of action at the time (i.e. there is no good and evil for the machine, at least not in the way we understand it). If it does not, it might solve all the problems of humanity. There's just no way to tell what will happen IMO.

Here's a link that I recommend checking out on AI:

https://news.ycombinator.com/item?id=8391896

It received little attention last time but to me it seems like a good piece.

View on HN

The worst case scenario is that AI is an all powerful psychopath - like the way extremist liberals ala 'the corporation' evaluate big business. We run our lives under the assumption that the rules and organizations we structure our societies around are something fundamental about humans that is 'good' or at least 'neutral'. If AI starts creating techniques of which to modify structures I already find incomprehensible, then it becomes a tool of which to analyze those structures as well. We know the present to be distinguishable from our history through differentiation. I agree it's moot. We can't know for better, or worse.

The message should be: We need more funding and promotion of AI safety research.

He, this remembered me the Portal 2 AI safety tags:

http://io9.com/5897036/the-rise-of-the-machines-as-told-by-w...

I understand that a singularity by definition can never be reached. It's like the speed of light or absolute zero; infinity just out of reach no matter how much energy you spend trying to get there.

A black hole is a singularity. You can reach it, you just can't come back from it.

Didn't Hawking a while ago say Black Holes don't exist?

Black holes most certainly exist. What he said is that their is a problem with the way we comprehend them now and that our current model for an event horizon may be false.

Does that not depend on your relative point of view?

Yes, if you looked at something going into a black hole, it would appear to you that it is slowing down exponentially the closer it gets to the event horizon, until a point where it appears to completely stop (before reaching it).

But in the aforementioned scenario, you are not an external observer but a part of the system going into the singularity (but analogies have their limits).

slowing down or speeding up? I could be mixing things up but I understand that if I was falling into a BH (lets imagine I'm infinity tough and not torn to shreds due to tidal forces) the universe would appear to be slowing-down/getting-smaller until it had dimensions of zero (or to put it another way, time would "stop") at C. An outside observer would see me moving faster and faster and "stretching" longer and longer.

SR makes my brian hurt. :P

First you have to understand that a black hole in fact, doesn't have anything really special (until you enter the event horizon), except that their escape velocity is greater than the speed of light.

But gravity applies to it like for any other thing.

So let's say you are the object going into the blackhole, and I'm looking at you.

From your point of view, the closer you get to it, the faster you will travel, just like when you get closer to earth, because it's gravity is stronger and stronger. So you will end up colliding with it (entering the event horizon) in a finite time (where time, is the time YOU feel, your proper time).

But a black hole is so massive, that space and time get stretched to extreme extents. So from my point of view, you'll be speeding up into the blackhole until a certain point, where you'll appear to be slowing down and stretching (the point where your increase in speed get smaller than the stretching of space-time), and the dimmer you'll get (because of redshifting). So i would see you going slower and slower and appear dimmer and dimmer until almost not moving at all, up to a point where I would simply not see you anymore (before you reached the event horizon).

All this also apply to other object like earth, It would appear to an astronaut that time on earth is flowing more slowly than it's own time (in reality it is of course unnoticeable , but it is true).

Also, to clear things up, you are mixing two point of view here, so here is the separation (you are the object going into the black hole):

"the universe would appear to be slowing-down/getting-smaller until it had dimensions of zero" From your point of view it would appear to be getting smaller, although I don't think it would go as far as dimension 0.

"(or to put it another way, time would "stop")" It would appear to the external observer, that YOUR time had stopped. But from your point of view it did not stop, nor did their time from their point of view.

"An outside observer would see me moving faster and faster and "stretching" longer and longer." You would feel you are falling faster and faster. The external observer would see you stretching longer and longer. (Although you would also feel like you are stretching, because of the difference of gravity between your head and your feet, but much less than for the external observer)

Ah I see where I mixed things up.

Although I understand that at C "time stops". But since time and space are the same thing, this is the same thing as "all distances are zero". From the point of view of a photon the Universe has dimensions of zero. This is why FTL is impossible; it's not possible to travel a distance less than zero; it's completely nonsensical. If we consider C a "speed" in the classical "speed of sound" the idea of FTL makes sense. But when considering C as a fundamental limit of space-time FTL makes no sense at all.

Another thing that might interest you:

"lets imagine I'm infinity tough and not torn to shreds due to tidal forces"

In fact you don't need to be infinitely tough. If you choose a black hole large enough (like the massive black hole at the center of our galaxy), you would just need a space suit and you could go into the event horizon without being pulled apart by tidal forces. You would be absolutly fine (their may be other dangers though...).

When talking about AI or technology in general the singularity isn't a mathematical one. John von Neumann who used it in this context first said:

> "ever accelerating progress of technology and changes in the mode of human life, which gives the appearance of approaching some essential singularity in the history of the race beyond which human affairs, as we know them, could not continue"

It's a different use of the word singularity outside the mathematical sense. Also absolute zero isn't so absolute it seems: http://www.nature.com/news/quantum-gas-goes-below-absolute-z...

Ah that makes more sense. He seemed to mean it as we use Stone Age, Bronze Age, etc. After we got bronze, we could not continue with stone alone.

That below-absolute-zero is in the quantum sense, not the classical sense. Entropy cannot be stopped (absolute zero) in the classical sense.

You are confusing a "mathematical singularity" with "the AI singularity".

In math, a singularity is an unusual point, not sure if it can't be reached. I think it's used in different areas to describe different things. For example, if you model water in a channel, after some time the waves accumulate and you get a vertical water profile, the depth is not more a function, the mathematical model collapse, but in real life the water just make some nice crashing effect.

How would an AI exactly exterminate humans? Are we talking about it physically murdering us or just making us so depressed as to our comparative lack of intelligence that we kill ourselves?

Like others suggested, there may be a point in its development where we think we contain it (in a lab, whatever) that is isolated by normal security standards.

And then at some point AI breaks all our digital defenses (because it evolved in a way we never expected) and proceeds to spread (through an Internet or other network which will be far more advanced than what we have now) wreaking havoc along its way. Isolated systems? Not a problem if we figure out a way to transport ourselves through some media that's never been used before (dust particles? or maybe plain old air?).

We can't protect ourselves from something we don't understand.

Sounds like a sci-fi plot but to me it has no evident problem that would make it impossible.

So it relies on a not yet understood law of physics. I don't think it's in any way reasonable to fear or plan our actions around not yet discovered physical laws.

It may rely on them or it may not, you never know. It may simply outsmart or trick humans, no unknown physics involved :)

It's not reasonable, I agree, it's flip-a-coin game. There's no point saying that it's gonna end bad, there exists one's opinion though that it's not worth the risk.

Mine is that it's fully worth the risk.

Reply on news.ycombinator.com