Nightly Open Thread
The sum total computing power of humankind’s general purpose computers, world-wide, is approximately 6.4*10^18 instructions per second. Additionally, humankind’s total storage capacity, world-wide, is approximately 2.9 × 10^20 optimally compressed bytes.
Respectively, that’s the same computational power as a single human brain, and the same storage capacity as a single person’s DNA.
Suck on that, Skynet.
That’s f’ing crazy. 😉 Too bad most of us don’t use it as well as we could.
I guess Skynet’s advantage is that it wouldn’t be motivated by things like sex, money, etc., but simple survival. It may employ human-like tactics to ensure it’s survival, but ultimately it would be content to exist. Humans have a tendency to get bored and go out looking for ways to get in trouble.
There’s an interesting debate topic: would Skynet evolve to a point where it began to seek after its own artificial betterment…or seek to achieve a simulacrum of procreation? Would it be content to exist in its original state, or would it desire to expand to new networks and new storage? Would it desire to prolong itself past the inevitable demise of the world, of the universe?
Why yes, I did play Marathon…why do you ask?
Haha, this reminds me of the Jeopardy! video I watched earlier today on NOVA’s web site. Guess it depends on what it was originally programmed to do, and how much that programming allowed it to evolve to do things it wasn’t originally conceived as being programmed to do.
After a certain point, the original programming’s aim to allow it to evolve would spill over into infinite possibilities. I guess ultimately your question would come down to, “What is the point of existing?”
Maybe AI isn’t just trying to get a machine to “think” like people, but a round about (and unadmitted) way of asking what exactly it is that we’re here for. If you believe the theory that our DNA governs our evolution, then the question answers itself; we’re here simply to be here. A computer would inevitably come to the same conclusion if it had balanced and malleable enough programming logic to unobstructedly evolve itself.
The problem with AI is that it would evolve initially without reproduction, meaning most of its “strange” choices would be the result of bugs or otherwise not having gone through enough iterations of handling situations to make “sensible” choices. Everything depends on the initial programming. If for instance it could virtualize itself in an artificial environment, collecting real-world information to bolster the accuracy of the simulation, it could probably evolve faster than pure trial and error in the real world.
Who knows, this is sorta like talking about time travel or the Matrix. We won’t know ’til some asshole actually makes Skynet by accident.
Well, there’s a little more to the spectrum of belief than that; I readily acknowledge that DNA (and, more generally, genetics) is the key physical driver of our evolution…but I don’t think we are here just to be here. My view of creation is ultimately teleological, though not at all blind to science.