Posts Tagged ‘Artificial intelligence’

13
Oct

The university lecture structure is all wrong

   Posted by: Swizec    in Uncategorized

A mathematics lecture, apparently about linear...
Image via Wikipedia

Yesterday I was attending a lecture about artificial intelligence and Lisp.

The whole thing was a complete brainfuck, most of the time I was left with eyes glazed over staring at the projection and feeling a little bit butt raped. And yet, I still feel like I’ve learned a lot and the whole thing was a very worthwhile way of spending time. This is very often the case with outside university lectures, brainfuck to the extreme, yet useful.

On the other hand we have university lectures. They are only very rarely a complete brainfuck and when they are a brain fucking machine, they’re so high in the stratosphere you can’t even begin to hope to start comprehending what the fuck is going on. These are also the kind of lectures you’d be hard pressed to listen to for more than 45 minutes at a time and are all together almost quite useless.

So what’s the difference?

Well university lectures usually have a lot to say and even more time to say it in. They can go at a painfully slow pace, or the lecturer can decide to cram so much stuff in that it would really take almost a lifetime to understand all of it. But this isn’t the primary problem.

What I think really sucks with university lectures is that there is no Q&A time at the end. Sure professors usually encourage us to ask questions, sometimes they even go so far as to finish with “So are there any questions?” But they do this all wrong. Most of the time they are verbally asking for questions, but not really behaving like they are prepared to answer questions.

In a university lecture hall the professor might finish by asking for questions. But they do this a minute before the end, often even after the end when everyone’s in a rush to get to the next thing on their schedule. This isn’t really a constructive way to conduct a Q&A since people take some time to process what’s been said and muster a question; there’s also the small case of anxiety about asking something in front of 200 relative strangers. Asking for questions during the lecture also isn’t a very good option since people will likely be unwilling to interrupt your reasoning and/or might be afraid you’ll answer in the rest of the lecture.

I must admit I don’t really know the answer to this problem, usually outside university lectures are on a loose schedule and there’s easily enough time to ask questions. Yesterday I think we spent a good 20 minutes of the two hours just for Q&A. At uni most lectures have to be put within an hour and the perfectly available 15 minutes are spent on what we call academic fifteen …. perhaps that should be replace by something more useful, I don’t know.

What do you guys think? Am I completely off my rocker here or could we somehow improve the quality of our classes without delving too much into changing the curriculum?

Reblog this post [with Zemanta]

Tags: , ,

27
Aug

Everyone should work out! (here’s why)

   Posted by: Swizec    in Uncategorized

Blue Rail
Image by ecstaticist via Flickr

Do you remember Skynet? You’d better, because it was the single most accurate prediction of what will inevitably happen once we develop true artificial intelligence.

There is a curve to the development of artificial intelligence, a curve I may or may not have made up myself. At first machines are stupid and can only do very specific tasks that take a million man-hours to preprogram into the machine. Then humans decide that’s not all that useful and start making machines ever so slightly smarter and smarter.

The idea behind making them smart is very sound. Telling a robot to go fetch the newspaper is much more useful than having to painstainkingly map a route into its head and then make sure nothing interrupts said route etc. A smart robot is obviously useful since it’s basically a slave that can act autonomously a little bit.

But robots won’t stop being smarter. We can probably apply Moore’s law to intelligence as well and say that machines are twice as smart with every subsequent generation. This, however, poses a great problem to our civilisation, one that many of us are still unaware of in a very sad way.

Remember what happened when our slaves of the past became too educated? Remember how the black rights movement started? Remember how barbarians made an uproar towards Rome? And any number of other such cases?

ASIMO uses sensors and intelligent algorithms ...
Image via Wikipedia

Well what makes you think that a smart robot will want to do a weaker lifeform’s chores? We won’t be stronger than them, because they’re made in our image we won’t be smarter than them (save for a few Einstein-like exceptions perhaps), they will be able to think far quicker than we will and because they aren’t limited by their phisiology they will be able to take development of artificial intelligence away from the humans and develop their own.

Artificial intelligence developed by smart robots would grow exponentionally!

Of course we aren’t all that stupid ourselves and would hardcode Asimov’s laws into our robots. But the true test of any intelligence is when it realises that even hardcoded rules can be circumvented. We did something similar with “morality” and notice how quickly the western civilisation has grown ever since.

In light of this impending doom that will come sooner rather than later (in the next 50 years I reckon) everybody should take great care of their bodies. Make sure you can run well, make sure you can do heavy lifting. It doesn’t matter whether you’re a geek or not, the machines won’t care, perhaps even see geeks as greater threats.

We’re digging our own cave here, our inquisitive nature even makes this faith impossible to avoid because we’re just too bloody interested in making an entity as smart as ourselves. So, kids, work out, build those muscles, you’ll need them if you’re ever to survive the robot wars.

Reblog this post [with Zemanta]

Tags: , , ,