Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Jimmy]

A technological singularity event would fit the bill for world domination, provided it's robot overlords you're wanting. Current estimates put it at somewhere around 40 years, so quite possibly within living memory. Google and other major IT players are already utilizing code that creates self-teaching machines, and combined with continuing developments in quantum computing, we're reaching that paradigm shift rather quickly.

[Teneb]

A technological singularity event would fit the bill for world domination, provided it's robot overlords you're wanting. Current estimates put it at somewhere around 40 years, so quite possibly within living memory. Google and other major IT players are already utilizing code that creates self-teaching machines, and combined with continuing developments in quantum computing, we're reaching that paradigm shift rather quickly.

All hail our new mechanical overlords? Beep Boop

[NullForceOmega]

There is a distinct lack of the requisite tanks and guns in the pitches I'm seeing here.  How can you overcome the forces of liberation and nationalism who will oppose your glorious unification attempt without tanks and guns?

[birdy51]

1). Acquire Nuclear Assets and a rudimentary space program.

2). Launch Nuclear Assets at the Moon.

3). Hold Moon hostage, lest world not give you the power you deserve.



Profit.

[chaotic skies]

There is a distinct lack of the requisite tanks and guns in the pitches I'm seeing here.  How can you overcome the forces of liberation and nationalism who will oppose your glorious unification attempt without tanks and guns?

Whoe said we had to overcome anyone? We could go the easier route, and convert/replace everyone to suit our needs

[Bauglir]

A technological singularity event would fit the bill for world domination, provided it's robot overlords you're wanting. Current estimates put it at somewhere around 40 years, so quite possibly within living memory. Google and other major IT players are already utilizing code that creates self-teaching machines, and combined with continuing developments in quantum computing, we're reaching that paradigm shift rather quickly.

Not sure. A singularity in the sense of "technological advance too fast for the human mind to keep pace with" will probably happen, and maybe within the lifetimes of currently-living humans, but a singularity in the sense of "recursive self-improvement without bound" won't because physics doesn't allow that "without bound" part. What people like to forget about hyper-intelligent AIs is that they're still constrained by computing power. No matter how clever the algorithm, sooner or later a physical machine has to execute instructions.

There are limits on how fast things can think, and I think they're as rock-solid as limits on how fast things can move, though less easily measured or calculated.

[chaotic skies]

What most people seem to assume about the whole "Self-learning AI that learns faster and faster, forever" idea is that at some point, the machine will start building more computing power for itself. Which, unless it is programmed to do so, it most likely will not before it runs out of computing power to use.

[Jimmy]

Yeah, most singularity event science fiction usually includes a point where the self-aware machine realizes the physical limitations of its current form and devotes itself to developing a more efficient storage mechanism or expanding itself laterally. For a decent Hollywood adaptation of the concept, I recommend the movie Transcendence (2014). It features AI driven by a human consciousness, but the principle demonstration of the impact of a technological singularity event is still relevant.

[Bauglir]

There's still only so much matter around that can be repurposed. I assume, but cannot prove, that the sorts of magical powers often ascribed to superintelligent AIs would require so much sheer mass to be obtained, especially since the beginnings will be especially crude, that somebody will notice long before the thing vibrates some electrons in its power cord in a way that creates a wireless signal, which it uses to connect to the Internet and break into laboratory and manufacturing systems and start cranking out nanobots that kill us all in one coordinated attack with no chance of resistance. Surely, the chunk of computronium the size of NYC will be spotted.

More to the point, there are also scaling factors on how much hardware you can glob together to work on the same task. There's overhead on coordinating and reassembling the subproblem solutions and so on, which is why you don't get 2x performance by assigning 2 cores to do something you used to have 1 do, even if it is something that can be neatly done in parallel. Signal propagation delays also become relevant at large enough scales. So you can't just pour raw processing power into making more processing power and expect your returns to be anything but diminishing.

This is why we have bureaucracies.

We're still low on the sigmoid, but it's not really exponential in the long run, is the point I want to make.

[TheDarkStar]

There are limits on how fast things can think, and I think they're as rock-solid as limits on how fast things can move, though less easily measured or calculated.

It's a branch of mathematics, actually.

[Bauglir]

It's a branch of mathematics, actually.

Yeah. The tricky part I was dodging was secretly the part where I don't have a good definition of what "thinking" means. The mathematics can establish a pretty good upper limit for information density, but that's probably not the equivalent number for "Maximum think capacity" that c is for "Maximum move capacity". I do appreciate the link though!

[Jimmy]

At a certain point, the theoretical limit on computational power becomes academic. Remember, we're comparing AI vs. the human brain. By the time any artificial intelligence begins reaching its limitations, it will already have an order of magnitude more processing power than any sum total of human minds, able to calculate and filter information at incredible speed with relentless precision, free of any distraction caused by biological needs such as sleep. Humans will be obsolete far sooner than AI will reach the maximum speed at which it can operate.

[Jimmy]

Science fiction writers have toyed with that idea for decades, however emotional response is still so poorly understood it's very likely impossible to replicate it meaningfully in an artificial intelligence. Our emotions are tied to both instinctual survival processes such as the need to procreate and defend ourselves, as well as more abstract concepts that relate to art and beauty. In contrast, these are unlikely to evolve naturally in an artificial system. Instead, we have models such as Asimov's Three Laws of Robotics, or the more recent work by organizations such as the Machine Intelligence Research Institute.

[Bauglir]

At a certain point, the theoretical limit on computational power becomes academic. Remember, we're comparing AI vs. the human brain. By the time any artificial intelligence begins reaching its limitations, it will already have an order of magnitude more processing power than any sum total of human minds, able to calculate and filter information at incredible speed with relentless precision, free of any distraction caused by biological needs such as sleep. Humans will be obsolete far sooner than AI will reach the maximum speed at which it can operate.

Sure, and that's the argument I originally made, if you'll notice. Technology accelerating faster than humans can keep up with? Certainly. But what I mean to curb is the notion of the infinitely clever, all-powerful intelligence in a box that has those attributes because you can contrive a physically possible scenario that allows any particular goal to be achieved. Especially, and this part is definitely a tangent, the kind that sneaks up on us because a machine designed to get better at, say, signing greeting cards went rogue.

"Vibrating electrons in its power cable to produce a wireless signal" is an actual strategy I've heard people mention is a possibility for an AI trying to escape confinement, and it's kind of the example I like to pick on. No meaningful confinement will have the processing power necessary to do that, if it's even possible in principle for the thing to control its power consumption that finely without being designed for it. The argument is akin to a prisoner saying, "Escaping will be simple. Humans can make keys, so I'll simply make the key to the cell door!" And that holds more generally for many of the wonders of tomorrow that tend to be associated with the singularity.

[Jimmy]

I can't recommend enough reading the short story The Last Question by Isaac Asimov for a great example of the theoretical limits an artificial intelligence might reach. It's even more impressive that it was published in 1956, six decades ago. In short, with exponential growth in theoretical intelligence, any perceived physical limitations become meaningless once undiscovered methods of storage and access become available through scientific advancement.

To use the key analogy, if you were to give your hypothetical prisoner the tools and materials required to manufacture their own key, it's simply a matter of time before they escape. Currently, the most obvious use for artificial intelligence is to sift massive amounts of data and analyse it. In fact, Google's Search and AdWords program are both run by DeepMind, a fledgling AI program. There's little to no chance of preventing AI from accessing the internet, and when that happens it has all the tools it needs at its disposal to effect an escape from whatever confines we place on it, short of completely shutting down the world wide web.