Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!

 



Forgot your password?
typodupeerror
DEAL: For $25 - Add A Second Phone Number To Your Smartphone for life! Use promo code SLASHDOT25. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. Check out the new SourceForge HTML5 Internet speed test! ×

Comment Re: Well it's easy to show superhuman AI is a myth (Score 1) 232

The human population is composed of experts, with divisions of labor. It is not unreasonable for AI programs to have areas of expertise.

In fact this is not true. The human population is composed of experts, some of whom have required in addition specialized skills due to division of labor.

Comment Re:Well it's easy to show superhuman AI is a myth. (Score 1) 232

Oh, and here's another example I just thought of.

I once read a book by an early aviator on techniques of navigating by landmark from the air. He recounts a number of feats of navigation by what were then called "primitive people", including one account of preparations for raid by a group of 19th Century teenage Apaches on an enemy village. None of the boys had ever been there, and so they sought out an elderly man who'd been there once when he was a boy. He described all the landmarks along the way, e..g. turn south at the hill that looks like such and so, a process took almost two days because the village in question was almost five hundred miles distant.

Now if a 19th C Apache had devised an intelligence test, chances are you or I would score retarded. There's no way I could give turn-by-turn directions to a place I'd visited just once thirty years ago. And if I could the chance you could just hear them and then go there without any difficulty is nil. We are simply too unfamiliar and unpracticed a task that is second nature to them.

Comment Re:Well it's easy to show superhuman AI is a myth. (Score 1) 232

You have to look again at how the tests are devised. Let's say you just invented an intelligence test. How do you know it's any good? You give it to a bunch of people and see if it confirms what you already believe about those people.

This culturally biases the tests in several ways. Let's say your test evaluates verbal and spatial mental performance. Naturally the verbal part will be biased towards not only native speakers of your language, but native speakers of your dialect. Then how do you weight verbal vs. spatial in your net socre? That's a cultural assumption. Even if you decide to weight them equally, that's still a weighting and represents a de facto judgment that one is not necessarily more indicative of intelligence than the other.

Then there's the stuff you don't include in your test, for example social reasoning and perception. Inferring other peoples' mental states and intentions is an extremely important aspect of intelligence, but it is also intrinsically culturally specific. Let's say you ask your neighbor whether you can borrow his car and he tells you it's broken. You know it's not broken. What can you infer from that? It depends on where you live. In the US you'd take it as a sign of disrespect, but in some cultures you'd infer that it would be inconvenient for him to loan you his car. Social perception and reasoning is one of the most important aspects of intelligence, but it is nearly inpossible to get a culturally unbiased mesaure of that.

Comment Re:Well it's easy to show superhuman AI is a myth. (Score 1) 232

in the middle range, 90 - 110 points,

IQ tests are also unreliable at the tail ends, for epistemological reasons.

How do you construct an intelligence test? You start with a collection of reasonable-seeming tests and you have a sample population perform them. You then rank them on test performance and assess whether your ranking confirms your preconceptions. So here's the problem with the tail ends: it's really hard to get a large enough sample of subjects to test the predictive value of your test with people who score three or more standard deviations away from the mean.

So while you can probably make predictions about differences in accomplishments between someone who scores 90 on IQ and someone who scores 110, I don't think you can predict much from a difference in IQ between 150 and 170, other than that people with an IQ of 170 will likely consistently score higher on an IQ test.

Comment Re:Well it's easy to show superhuman AI is a myth. (Score 1) 232

You seem to be confusing the concept of intelligence with *measuring* intelligence.

You know you're right. But I think there's a good reason for this: magnitude is an intrinsic part of the concept. I've never heard anyone talk about intelligence except as a concept that allows you to rank things (e.g. Chimps are more intelligent than dogs, which are more intelligent than gerbils). So to apply it to an entity like a human or a program is to implicitly measure that thing.

What I'm saying is that the concept is useful but of intrinsically limited in precision.

Comment Well it's easy to show superhuman AI is a myth. (Score 3, Insightful) 232

Because intelligence as a single-dimensioned parameter is a myth.

We already of have software with super-human information processing capabilities; and we're constantly adding more kinds of software that outperforms humans in specific tasks. Ultimately we'll have AIs that are as versatile has humans too. But "just as versatile" doesn't mean "good at the same things".

So it's probably true that software is getting smarter at exponential rates (and humans aren't getting smarter as far as I can see), but only in certain ways.

Comment Re: Bullshit. (Score 2) 128

"Using a chat program to hide " doesn't even make logical sense.

It does if the chat program using public key encryption between the users. In that case even the mediating servers don't have access to message contents.

The scheme is flawless -- but then it almost always is unless it's devised by a total ignoramus. What they get you on is implementation.

Comment Re:You can't generalize. (Score 1) 387

It does *sound* a bit sociopathic, doesn't it? But sociopathy is a pathological disregard for the rights of others. While deception is often used to violate someone's rights, but it can *also* be used to protect someone's rights.

For example if I knew an employee was embezzling money, I don't have to tell him I know. I can deceive him into thinking I'm not on to him until I gather enough proof or discover who his accomplices are. This is deceptive, but not a violation of his rights.

Comment You can't generalize. (Score 2) 387

Anyone who works on unauthorized personal projects should certainly expect to be subject to firing. But as a supervisor I would make the decision to fire based on what is best for my employer. That depends on a lot of things.

I don't believe in automatic zero tolerance responses. The question for me is whether the company better off booting this guy or disciplining him. Note this intrinsically unfair. Alice is a whiz who gets all of her work done on time and to top quality standards. Bob is a mediocre performer who is easily replaced. So Alice gets a strong talking to and Bob gets the heave-ho, which is unfair to Bob because Alice did exactly the same thing.

But there's a kind of meta-fairness to it. Stray off the straight and narrow and you subject yourself to arbitrary, self-interested reactions.

Now as to Alice, I would (a) remind her that anything she creates on company time belongs to the company (even if we're doing open source -- we get to choose whether the thing is distributed) and (b) that any revenue she derives from it rightly belongs to the company. But again there's no general rule other than maximize the interests of the company. I'll probably insist she shut down the project immediately and turn everything over to the company, but not necessarily. I might choose to turn a blind eye. Or maybe even turn a blind eye until Alice delivers on her big project, then fire her and sue her for the side project revenues if I thought we didn't need her any longer. If loyalty is a two-way street, so is betrayal.

Sure, you may rationalize working on a side project as somehow justified by the fact your employer doesn't pay you what you're really worth, but the grown-up response to that is to find a better job; if you can't, by definition in a market economy you are getting paid at least what you're worth. If you decide to proceed by duplicity, you can't expect kindness or understanding unless you can compel it.

Comment Re:60Ghz (Score 1) 140

I agree it sounds impractical. So I looked at the patent -- which not being a radio engineer it's perfectly safe for me to do (n.b. -- it's always dangerous to look at what might be bullshit patents in your field because you open yourself up to increased damages for using common sense). But I was a ham radio operator when I was a kid so I do know the lingo.

There are a number of problems with broadcasting power, starting with the fact that it's inefficient to saturate ambient space with enough radiation to be usefully harvested. But that's not what they're proposing. 802.11 ad operates in the extreme microwave range -- about 5cm wavelength aka the "V" band. This band is also unregulated so you can try weird things in it. What they propose is to use an array of antennas to create a steerable beam -- like a phased array radar. This would confine the power to a specific plane so that you wouldn't have to saturate all of ambient space with power. The beam steering would be done "dynamically", which I take to mean it would figure out how to maximize signal strength with some kind of stochastic algorithm. So it might not work if you are unicycling around the room.

And because the wavelength is so short an antenna array would be relatively compact.

Even so, it doesn't sound that practical. It's bound to be limited to line of sight, for example: the V band does not penetrate walls or the human body at all, in contrast with the S band that conventional wifi operates on. I can certainly imagine applications for it, but making it practical for charging your phone is apt to be very expensive. You'd have surround yourself with V band antenna arrays.

By the way, reading this patent reminds me of why I hate reading patents. They're infuriatingly vague in order to make the claims as broad as possible, and yet are cluttered with inanely obvious details ("the radio receiver can include active and passive components") and irrelevancies (the device may include a touch screen). I think the purpose may be that someone trying to figure out whether the vague language applies to a cell phone will think, "I don't know WTF this is claiming, but look this phone *does* have a touch screen." It just shows how broken our patent system is.

Slashdot Top Deals

"There is no statute of limitations on stupidity." -- Randomly produced by a computer program called Markov3.

Working...