Faint

It would be nice to know something about the accuracy of the age estimate (stdev? or 66% of 29-year old's trials are between blaah and blaah?), and also, how many tries did the test subjects get. I got a couple of "we are not sure you are human" ratings after a few tries (around ten tries total).

Not all birds are short lived... Over 60 and still producing offspring:

https://en.wikipedia.org/wiki/Wisdom_%28albatross%29

White has first-mover advantage, maybe your browser does not support asm.js? At least on Firefox blue won, but the page load so slow I don't care to repeat the experiment right now.

Well, if it's actually a Dyson swarm in the making, we'll find out, watching the star go gradually dark during the next few hundred (or thousand?) years..

You don't have have to put all of the 52MWh side-by-side.

For purposes of deciding where money and resources are best spent, I don't think it makes any difference whether there was someone with intent to harm or not.

You can spend money to construct safer railway crossings, you can spend money to hire more cops or NSA operatives, you can spend money to educate cops, or you can make rule that all cars must have safety devices that stop them if they intend to cross railway when a train is coming, which will make cars more expensive, and so on and so on.

Never the less to prioritize these, statistics are important.

Confidence, or conscious willingness to take steps to unknown (will to explore)?

Pigs age about 5 times faster than humans. Wouldn't pig organs age 5 times faster than humans also? That might be one more thing you needed to tinker in pig DNA. Make pigs (or at least the target organs) long lived.

All this tinkering begs the question: Which is easier, tinker pig DNA to make pigs compatible with humans, or tinker human DNA to make humans with pig-like-brains that grow fast enough to be used as organ source? Does it ethically make any difference?

What if we could grow human bodies with no consciousness what so ever (this might not be even super hard, or require DNA tinkering, just damage the fetus where appropriate brain parts are about to develop), wouldn't that be the most humane option? Since we wouldn't be killing anything, but hunk of human meat that never had a conscious thought...

1.4 billion people, just have to nitpick on that.

Drag force is proportional to 2nd power of velocity, power lost to drag is force*speed, which is 3rd power of velocity.

Air drag is proportional to the front facing area of the vehicle, that is 2nd power of cyclists dimension (for example, height, we are assuming that humans have approximately similar shape, regardless of size). Max power grows to 3rd power of cyclists dimension, because it's dependent on the cyclists muscle mass. Maximum aerobic power grows somewhere between 2nd and 3rd power, because of the fractal shape of lungs and veins (for example, think about lung surface area, the smallest folds in lung have same dimension in bigger and smaller guys, so the area grows faster than 2nd power of cyclist's dimension).

It then follows that bigger guys do well pushing against air (since frontal area grows slower than maximum aerodynamic capability), and smaller guys do better dragging themselves uphill (since mass grows faster aerodynamic capability). You can easily find practical examples of this: say Fabian Cancellara, who has won time trial world championships 3 times + almost anything else riding in fairly flat ground, including several tour prologues and stages before getting into mountains... but when you get to the mountains, it's the featherweights that rule, so likes of Cancellara can't ever win the tour.

So... to propel something as fast as possible against air, you need big and strong cyclist (with huge lungs to match), just like the guy in the video looks like. Power/weight has little to do with it, since air drag is most of the resistance, and they can accelerate for 5 miles before measuring speed (you would need power/weight if you needed to accelerate fast, which is not the case).

Looks cool, but would have been a lot smoother in webGL

Well, you'd end up with something like this:

https://www.mtholyoke.edu/~wang30y/csp/ParabolicDish.html

GaAs panels have ~40% efficiency. It's not easy/dirt cheap (you need high temperatures, and materials to go with that) to get that much efficiency out of steam turbine (or Stirling engine). For example your typical nuclear power plant would have about 33% thermal efficiency. And with the panels you have less moving mechanical parts that need servicing. And the panels are silent vs. steam turbine. Lighter also. Pure solar thermal has been tried/used many times before.

They are the geostationary satellites (1 orbit takes 24h)

Generally causation and correlation can only be distinguished by running experiments. Given that we only have 1 earth climate, long time needed to gather reliable results, and the quite apparent correlation, which experiment would you like to run for next 50 years, slow down co2 output to see if climate warming slows down, or continue increasing co2 output, and see where that leads?

Infact, increasing co2 output would be kind of boring experiment, we've tried that already.

I've heard a couple of other sites like this, is there a googlable term for them? A directory? Soudscape site? The rain sound loops a bit too often in longer listening.

Maybe it's because the sheep make sound when they start running, and the OP test is visual only. I feel that I react faster to sound... I dunno tho.

If people (politicians, and public) would just get rid of their prejudices and we could just develop a nuclear reactor for outer system exploration and be done with it...

Fuel would not be such a pain to manufacture, there'd be oodles of more power available (more bandwidth for communications, nuclear electric propulsion for faster travel times), and the reactor would be just as safe, if not safer to lift to orbit (unused nuclear fuel is not radioactive, Pu-238 is, although it's so well sealed that it's practically impossible to free it to environment on an accident).

And it's not just oil. It's all the minerals we've been mining since the dawn of civilization - take copper: before copper age one could find nuggets of pure copper metal [1], in ancient Egypt it was produced from ores with 10-12% copper [2]. 21st century copper ore has an average of 0.6% copper in it [3]. I doubt that ancient Egyptians could have done much anything with 0.6% ores.

[1] http://en.wikipedia.org/wiki/Metallurgy_during_the_Copper_Ag...

[2] http://www.reshafim.org.il/ad/egypt/timelines/topics/mining....

[3] http://en.wikipedia.org/wiki/Copper_extraction_techniques

You probably mean static, not strong typing.

Denying tech sales to China, at this stage is just causing minor nuisance. In not-so-far future China is going to be much stronger economy than US ever was. If they want the bigger computers, they will have the bigger computers. Just calculate any "damage" done with those export rules on Chinese military - is it $2B more that needs to be spent in order to get the same capability? Or $20B? You can just watch that money float right past you to their domestic HPC and CPU industries. In effect, US is trying it's damnest to make China competitive in the one major high tech export industry that it still has undisputed world leadership in.

Only circumstance this would make any sense would be if you were sliding fast and inevitably to open conflict, and causing them setbacks worth of a few months to a couple of years would be vital. Since that is not the case, this has nothing but "strategic blunder" written all over it.

I think people might overestimate the effect of machines being capable of improving themselves.

I don't believe in "magic algorithm"s, that we can't find, that perform dramatically better, but computer can find. I think we are capable of taking, in a few years time, advantage of most hardware. Dramatic improvements in performance thus require improving physical matrix the computation happens in. That means the prospective machine intelligence needs to have access to all the infrastructure thats needed to design and produce such matrix. It also will need to make experiments, and do things by trial-and-error, just like we do, even tho it could catch a hint easier than we do.

Think of chess. Seeing a couple of moves further in to the game might require 1000 times more raw computation. Such power might guarantee victory over lesser equipped competition. It still doesn't mean you'd be even close to "solving" the game. What I mean is, "intelligence" is not the only requirement needed for self-improvement. In the end, it might not even be the bottleneck at all. For example, suppose we are planning the next generation of a transistor for some IC. Suppose a team of humans needs to do 100 experiments to get the process right. A perfectly smart AI might still need to run 90 experiments, no matter what, because it still needs to extract the same information that humans would from those experiments.

The machine needs to do all the chores (of R&D) that we already are doing. I'm not saying they couldn't be done better, but I'm saying they still need to be done, and any creature doing them still needs to spend the time and resources needed to do them. And it's not all brainwork. I have no doubts vanilla humans can be outcompeted at some point. But even when human performance has been exceeded, technology might not immediately take off rocket-like to a world we cannot comprehend.

Still, I wonder, what the hell are we all going to do when paying the power-bills and capital costs of AI become cheaper than hiring a human..

Actually, we have _not_ stopped evolving because of natural selection. A significant proportion of people choose not to have kids nowadays, or are just having one child. In a generation or five all this "I don't want kids because there are too many of us in this planet already" crap will be over, and we'll only have people left (in developed countries) that get drunk and don't use condoms and decide to have the kids anyway, people prone to religious belief systems that forbid contraception, and just people that plain have very strong ingrained urge to have a family/kids, despite career pressures and all the information/stimulus overload from internets and 24h connectivity and all.

Why would the batch auction need to be FIFO? What about something like this:

1) When opening a "new tick", release data about trades in last tick, and start accepting new orders. Information about orders submitted to this tick is kept secret. Allow setting order price to practically arbitrary accuracy. 2) After tick closes, match orders in _random order_. That is, it should not matter at all on at which point in time inside this tick order arrived 3) repeat

Allowing arbitrary accuracy would help turn the competition from speed or stuffing the order book with huge number of orders to "spoof" the random matching, towards competition of just providing lower spread. I understand that it would still advantageous to hold submitting trades until very end of the tick, to take into account as much outside information (other markets, etc.) as possible. To counter this, maybe the exact length of the tick could be randomized also. What do you think?

I think the most exciting opportunity here is actually compiling models to run in specialized inference hardware! Lyric semiconductor a.k.a. Lyric Labs of Analog Devices is working towards such goals (http://dimple.probprog.org/, http://arxiv.org/abs/1212.2991). I hope that they will get some hardware out at some point.

Also, doesn't it strike you strange that we are building simulations of stochastic phenomena (MCMC) using deterministically behaving components? What if we used stochastically behaving components as building blocks to begin with?

Most electronic components start to behave more or less stochastically when you try to run them with as little power as possible, and try to scale them down as small as possible. What if you could build MCMC simulator for the problem of your choice directly from stochastically behaving components? Just think of all the transistors you nowadays use just to generate a random number for simulating 90% probability of something.. they all are components that are manufactured to such tolerances, and run with such power levels that their probability of error is something like on the order of 1e-24 (for one computation). Doesn't that strike as a humongous overkill, when all you needed is something like "1, most of the time"..?

For more related cool stuff, google for imprecise hardware.

Could we make him pay a few bucks?

Specifically, can we make him traffic multiply? I wonder what exactly is he doing with request headers... maybe this could work:

1) set up page /fluffy with wildly compressing contents, say 50MB of $£€$£€$£€$£€$£€.. always force gzip encoding 2) set up a few bots (amazon?) to download that page from his site, but do not accept any compression

Start the attack on some time the guy is probably sleeping, it might go on for a few hours before he notices, costing him a couple of hundred bucks in bandwidth.

Or maybe just some cpu waste in same vein: the guy has to open the gzip before forwarding to do string replace and re-zip it afterwards, so you can make sure that the content REALLY balloons..