... from which heroin can be refined, earning billions of dollars a year (well, nobody knows the true number). Just one problem with that!

If there's lithium etc., then someone can pour vast amounts of money into Afghanistan without having to admit that they're either funding narcotics or enriching corrupt officials. Oh, and metals are less easy to smuggle than poppy syrup, which also gives the recognized government an advantage in trading. So it becomes possible -- hardly likely, but possible -- that Karzai or his successor will be able to afford a real army and stand up against the Taliban.

The place will still be run by warlords, just not by druglords. I don't know that I care much either way.

I forgot to say that, almost by definition, a solar sail is not very useful when you are far from the sun: the inverse square law works against you. Neither sunlight from behind you nor starlight from in front of you provides much thrust.

Even when you are near Earth orbit, sheer scale is a big problem. To propel a small spacecraft at useful acceleration, you need a sail the size of a small town (and I live in the USA, where small towns are not very small spatially). The sail also has to be very thin in order to save mass, perhaps a hundredth the thickness of the Japanese sail, so handling and controlling it will be tricky, as Clarke points out in _The Wind From The Sun_. He envisages the sail as one or a few huge panels; I wonder if many small panels would be easier, both to use and to manufacture. Oh, there are lots of problems to solve. And then you wouldn't (notwithstanding Clarke) use the sail for anything shorter than a trip to Mars.

Somebody will probably point out that if you could focus all the sunlight that falls on a small town, you could make a hostile spacecraft very uncomfortable, though you couldn't get them any hotter than the surface of the Sun.

How does it stop? If it accelerates tangentially to the Earth's orbit, which is still the most efficient way to get to another planet, then it can decelerate by tilting the sail the other way. In each case, the acceleration vector will have a component outwards from the Sun; the ways to cancel that include furling the sail and waiting for the Sun's gravity to do the job, using a nearby planet's gravity, aerobraking in a nearby planet's atmosphere, or lithobraking. If none of the above work, then perhaps you can't stop. A bizarre scheme that has been suggested would be to bring a second, smaller sail along and use it to collect light reflected from the main sail towards the Sun (you cut the main sail loose and let it drift ahead of you), thus providing reverse thrust until the main sail is too far away. Hard to be sure how well this would work.

Debris hitting the sail? A few pinholes will make no appreciable difference to its performance. A real sail would have to be made with some sort of "ripstop" reinforcement.

Max speed? You have a misconception here: solar sails don't use the solar wind (much), but the pressure of the Sun's light. Since e=mc2, momentum equals e/c. I don't have the formula handy, but the important factors are the thickness of the membrane (thinner is better) and how close to the Sun you start (closer is better, provided the membrane doesn't melt). In theory, solar escape speed is attainable, if you're only pulling a small payload. Significant fractions of the speed of light are not attainable.

Scooping up the gas would need one **** of a scoop!

This tells you what "going to college" really is for: making it clear to prospective employers that you are willing to put up with an awful lot of c**p because you badly want a job, a place in The System. Someone who dropped out of a college might well drop out of a job the same way, and thoughtful employers don't want that. After all, they spend your first year training you, whether they admit it or not, so they want to get some work out of you after that year.

Getting an MBA has been this way for a long time, according to friends of mine: they spent most of their time drinking with other MBA students and interviewing with potential employers. The employers mostly cared about the fact that the candidates had got themselves admitted to a Biz School, not about what (if anything) they learned there.

Most of the world, and especially the country that was so newsworthily stubborn at Copenhagen, has been keeping democracy on hold for the past several thousand years, but the globe is getting warmer anyway.

The definition of AI research, remember, is "spend six years writing a program that does badly something that a six-year-old can do well." Most of the things that are hard for AIs are things that humans do effortlessly. The things that humans spend years learning at school, like how to add fractions with coprime denominators or how to put commas before subordinate clauses or how to beat me at chess, can be done effortlessly by computers and nobody even calls it AI these days. So it should not be hard for an AI to learn them, once it can pass the six-year-old test.

The real question is which kind of AI you want. Do you want one that can do exactly what most humans can do? Or one that can win a Nobel prize but cannot make small talk about the weather? In other words, I do believe that step 2 is achievable without achieving step 1. After all, we already have AIs that can play grandmaster chess but cannot make small talk.

Since I'm too lazy to read TFA, can someone give us an idea which kind of AI it was talking about?

One of the questions the Augustine commission explicitly bucked to the politicians, rather than offering any advice of their own, was whether NASA would be a space program or a jobs program. I was very sure the politicians would duck that question. And I fully expected Obama to opt for a jobs program while pretending that wasn't what he was doing. So give the man a round of applause.

Someone will still go to the Moon, and I now can hope it'll be someone who knows why they want to go there!

It's not impossible. I even have an idea for how it might be done. Divide the country into thirteen geographic regions. (That would be for small countries; larger ones could go to as many as 50 regions.) Give each region its own capital city and allow the people who live in the region to elect a smallish government, which would have power over that region only. Maybe even allow these governments to send specially chosen officials to advise the national government, but no more than two per region, to avoid crowding. Lay down rules for what the national government can and can't force the regional governments to do. And (here's the kicker) forbid the regional governments to impose their own controls on immigration or emigration, so that anyone who didn't like the government in their region could transfer to one where they did like the government.

Pretty soon, some regions would have larger populations and their governments would have more money. Kinda like businesses that have developed a better product or service. Nifty, eh?

Now for the bad news: this has beep-all to do with NASA, and I'm sure someone will explain why it has no chance of working in real life.

The plants we eat nowadays are not just the result of selective breeding as you think of it, i.e. crossing one ancient variety with another. Since the time when it first became clear that ionising radiation and certain chemicals could cause mutations that might result in changes to the daughter plant, seed companies have been trying it, throwing away the ones that turned out toxic, and keeping the ones that seemed to give an advantage. [Citation needed, sorry, lost it for the moment.] The difference between this and modern genetic engineering is that back then they had no idea what changes they were making; they just rolled the dice.

So to repeat what others have said above: the final product should be tested for safety, not the process that produced it.

I'll add the other side to your story. (Full disclosure: a member of my family has worked in mental health, though without prescribing any drugs.) Both in people I've known personally and in case studies written by a doctor who believed that many of the worst mentally ill people are not "untreatable" despite the psychiatrists' opinions, it has been clear to me that meds sometimes provide a vital breathing space, so the patients can at least make good use of the hardware inside their heads, and thereby get themselves better, given a software expert (that is, a therapist) who can reverse engineer the bad software in their heads and help them uninstall it.

The problem seems to be that therapists who have enough imagination and perseverance (and, in extreme cases, enough physical courage) to help these patients, and enough respect for them to prescribe minimum rather than maximum amounts of medication, are scarce and expensive. You think drugs are expensive? They are, but the cost of a therapy session can pay for a lot of pills. So there tends to be a bias towards pills. Easier to whack the hardware with a hammer than investigate the software ... but running bad software on good hardware still isn't going to work -- and please don't take that personally! Plus it may take time, and money, to tell whether the problem is in the hardware or the software (or in both).

I have also heard stories, by a Stanford professor no less, where anti-psychotic meds were prescribed in "wanton" doses in mental institutions. So although I've never been through what you have, I am sure your stories are as real as mine. On the lighter side ... well, it wasn't light for him ... I once knew someone who suffered from depression, so his doctor had him sent home from work on (quite generous) disability pay; the only problem was that most of the things that made him feel good and could have lifted him out of depression were part of his job. Argghhh.

Brad Templeton says it better than I can: http://www.templetons.com/brad/transit-myth.html

I'm no more into knocking mass transit than he ... was ... but I can no more help manipulating numbers than I can help breathing, and the numbers show that mass transit works well where you have heavy population density, which most of the USA does not. It works even better when you have low to moderate income and low car ownership, which most of China still has.

And since you don't ask, no, I'm not hoping to impoverish the USA so that mass transit becomes the optimal choice. It'll happen anyway.

For that other activity you mention, am I right in guessing that your couch measures 78 by 60 inches and comes with head-to-toe continuous-coil steel springs, 200-thread-count satin sheets, and a down comforter?

Like you can't do online banking via touch-tone phone? I used to do that back in the Dark Ages; these days, even developing countries have mobile phone networks, where you can do your banking and market research by text message. Sorry, someone is just trying to scare us.

I remember, back in 1993, explaining Open Source to a neutral third party in the presence of a self-proclaimed Libertarian (not sure if he used a capital L ... writing had not yet been invented then, y'know). He heard me out, then launched into a diatribe about how Richard Stallman was a Bad Guy because, among other reasons, he was charging absurdly high prices for ports of GNU to certain hardware platforms.

When a Libertarian denies that a business has the right to charge any price it chooses for a product or service (and go out of business if nobody feels like paying that price, or make vast profits if enough people do feel like paying it), please give me some Socialists.

Actually, I wish I could stop my kids from singing at home. (They wish the same about me.)