Murderous machines to be the next big thing

In an article about artificial intelligence -- and Google's intention to develop and use this capability -- there is a particularly chilling note.

The authors speak about the ways the United States currently kills enemies by using drones. That's bad enough but it's about to get much, much nastier:

"Right now everything we have is remotely controlled and there's always a human in the loop," Akerson says. "We're heading in the direction to give the decision to kill to an algorithm."

That's right. The plan is to unleash murderous drones that don't need humans to push an actual kill switch. They'll be able to do it all by themselves.

Oh, brave new world that has no people in it. Scary stuff.

When (virtual) bots rule the Earth

Are we in danger? Will a super-intelligent A.I. take over the world? And if so, what will happen to humans? An interesting story about this prospect ends with the following:

Price is co-founding the project together with Cambridge professor of cosmology and astrophysics Martin Rees and Jann Tallinn, one of the founders of the internet phone service Skype.

Hmmm, so one of the guys looking into this is a founder of...Skype. And if you replace one letter in "Skype" and add another, you get...Skynet. (Gulp.)

And yes, I'm kidding. I would welcome our new, artificially intelligent overlord. Surely it would do a better job of managing human affairs than our elected leaders.

Intelligence not required

There's a story today on physorg about the mental abilities of bees. Here's an excerpt:

The team's findings suggest that bees are able to solve complex routing problems by learning, without needing a sophisticated cognitive representation of space. Dr Lihoreau explained: "Despite having tiny brains, bees effectively used gradual optimisation (comparing several different routes), to solve this famously complex routing problem which still baffles mathematicians 80 years after it was first posed."

A tiny bee brain that we wouldn't assume is capable of very much, and yet it can outpace us in this area. I love that the bees managed to learn and come up with this routing solution without a large brain. That's so cool.

I sometimes wonder if intelligence is the holy grail that we think it is. My favorite kind of sci-fi books are the ones about alternate, alien sorts of intelligence. Peter Watts' "Blindsight" went wild with this notion. If you enjoy the idea of varied concepts of consciousness, pick up a copy. The book even asks whether intelligence requires a self. Could there be another kind of intelligence, almost the opposite of a human's? His answer is fascinating.

Consciousness and intelligence are insanely interesting topics. Remember, folks: we don't even know how our brains create the illusion of "selves" -- or for that matter, what these "selves" are. It's just wild that we know so little about what we are.

We are truly an infant race. Here's hoping we reach adolescence.

One step closer to artificial intelligence

I love the concept of machine learning. It's how IBM's Watson learned to play Jeopardy. It simply looked at all the questions ever asked on Jeopardy, and read all the wrong and right answers, and from this information, figured out the pattern of "correct answers". I think that's mind-boggling.

There's another story out today about machine learning. The article at physorg explains how a computer system learned to play the computer game "Civilization" -- by reading the manual! It's an amazing story. Check it out. This field is really moving now. It's not quite time to welcome our overlords but machine learning is growing by leaps and bounds.

Like the Chinese, it comes. Zillions!