XKCD: It's more than a comic

I don't think that robots would (or should) reproduce uncontrollably. Intelligent robots, at least.

Let's assume that we create the robots to serve humans/humanity. We would program them so that they feel "happy" when they can serve humanity, and do all things to achieve this. They would of course need to reproduce and take care of their own self-preservation, but they would do this only to the extent that is beneficial to humans. Humans like to reproduce and compete for resources because these instincts are programmed into us by evolution. But a robot's "instincts" wouldn't be made by evolution, but by the programming of humans. Even if the end result would be a result of some kind of evolutionary algorithm, that would probably use a more sophisticated fitness function than simply the ability to stay alive. So ultimately, the robots would reproduce only to the extent that is really needed, thus spending more resources on delivering more resources to humans than on pure self-preservation.

Of course, it is an overly idealistic and simplified view. First of all, how do we define something as beneficial to humanity? How can we prevent the robots to be abused by humans for an evil purpose? These are questions I don't want to cover now, because they are way too difficult for me to answer. Asimov has some good reads on the topic.

Then, there is the problem of morality. When we create the first robots that will produce all the robots who will eventually govern us, ideally in a way to make us happy. These robots are programmed to follow a certain morality that dominates the society they were created in. But morality changes with history. People think differently in different ages. In a society governed by robots, however, they will force us to follow the morality of the age they were created it, thus preventing any future changes in society. Now, THAT is how you create a Crapsaccharine World.

I once read an paper that suggested the most important thing about making AI would be to make robots understand what things mean to humans.

Otherwise we'll hit situations where you ask the robot to do something and it misunderstands what you want because it's working from a non-human reference point.

I.E, asking the super AI to cure cancer and it decides killing everyone is the optimum solution because it doesn't have human values.

That reminds me of this xkcd.

More about the eclipse. I like the lady's answers. :) The reporter said "science" so many times, it lost all meaning. x_x

Scientists can be excited by things that have nothing to do with their direct pursuits, you silly reporter. They don't check their sense of wonder at the door.

Apparently the next Total solar eclipse in the US will be in 7 years, so not exactly once in a lifetime

I checked a site that listed the upcoming total eclipses. This one isn't even visible in Finland. You lucky bastards get to have several more after this in the next couple of decades, whereas we've only got about one or two in that time.

I'm not sure that "lucky" really comes into play when you consider that Finland has quite a bit less land area than the continental United States and is also located pretty far north, both reasons why it would see fewer eclipses.

Even compared to all of Europe, you're having more in the near future, so to that extent my comment stands. Both of your points about Finland are correct, though.

"with a sunset on every horizon"? I don't get that part.

You guys get the aurora borealis, though, so in terms of cool sky-related phenomena it kind of balances out.

edited 16th Aug '17 2:17:02 PM by Galadriel

The sky looks like sunset all around the horizon while the sun is obscured, I guess?

Here in Hungary there was one total eclipse in 1999. That was really fortunate, as the totality went right through the most frequented tourist attractions of the country. I also remember a few partial eclipses, one of which was about 1 or 2 years ago, was total in some parts of Europe.

As for scientists, they are still humans, who are just as fascinated by rare natural phenomena as other humans, naturally. If they happen to be astronomers, they are especially excited, but that's not just because of the eclipse (there is really nothing special about it in an astronomical point of view), but because they can study the Sun in ways they cannot do otherwise. is some more interesting information about this.

To me, it's more like "sunset all around the moon" while it's obscuring the sun.

If we ever create fully sapient AI we will just program them to obey the law, not Asimov's, the human laws we have on the books. If they are truly sapient, then they will be more than fully capable of handling contradictions and ambiguities exactly the same way we do- by guessing. If they are less than fully sapient, then we will have to provide a way for them to receive a full range of "If-Then" commands in response to environmental stimuli. The whole point of self-learning AI is that they generate their own "If-Then" response patterns using a pre-programed long-term goal state as their guideline. True, we can't know ahead of time what response patterns they will develop, that's the whole problem with the paperclip maximizer problem, but they won't violate their long-term goal. They won't even be capable of wanting to.

So it's a matter of trail and error, with our fingers on the off button.

#1878: Earth Orbital Diagram

My brain aches looking at this.

Don't look at that diagram too closely, it will hurt your eyes.

I love comics like these when you slowly try and start to pick out which ones are gibberish.

They're all basically gibberish, aren't they?

I don't know, I don't get half of them. That's the fun.

"Prolapse" is the one that made me go "wait a minute..."

I don't think any is legit. "Perihelix" seemed so at first, but it's actually "perihelion".

It also mixes up solstice and equinox.

Dimples of Venus is a legit

#1879: Eclipse Birds

Yes, nature is amazing. *scared*

Cue "The Birds" theme music.