If you are right, then this lawsuit is simply one of several steps it will take to make Azure (and others) a complete failure in the EU market. If by using Azure you're subjecting yourself to US jurisdiction and accept responsibility for any breaches of EU law that result from this... Bye bye US-owned cloud services.

Go ahead, shoot yourselves in the foot. The US is burninig goodwill like there's no tomorrow...

Yeah, been there, seen that. Yes, I googled for a stationary bike. Guess what, I bought it. No matter how many times you show me an ad, I'm not going to buy another (at least until this one breaks).

I do not have a Facebook account, but I'm sceptical about targeting. Not just because I'm allergic to stalking (I might be). Just because I've seen it and it doesn't work. I mean Youtube.

I have to clean some youtube cookies every couple of weeks. I view sufficiently different content, that the algorithms they use simply backfire. As my browser window shows more and more "chosen for you" videos, I see less and less things I might want to view. I'm beginning to thing I'm really that atypical, since the things they pick are usually completely uninteresting to me, although I can probably retrace the reasoning behind the choice.

I have no idea if FB ads work the same way, but the ability to set "do not target" seems nice. Not nicer than not having an account and periodic cookie cleansing though.

Which makes this story even more interesting. In this case you're not adding antioxidants to your body, which might protect you or harm you, evidence for both has been reported. You're adding it to the food while it's being prepared, reducing the amount of PAHs, which are doubtlessly bad for you. More antioxidants in your body - open question. Antioxidants acting on your food before you eat it, removing other dangerous chemicals - clearly a good idea.

Analogy: drinking base to reduce bad effects of ingested acid - a very bad idea that might sound good. But add the right amount to the acid you're forced to drink with a gun pointed at you before drinking - tada.wav, you're basically safe, the worst outcome is that the result will be slightly poisonous (unless you prefer acid burns to mild diarrhea).

Ok, Nash equilibrium, finally the correct approach. But here's an interesting variation, seemingly very similar, but a bit harder: same probabilities, but introduce a third side with a predefined algorithm and infinite budget. The game still looks the same to you as the player: 50% rock, 50% human RPS player, but in fact suddenly the game depends on so many details...

The setting: you play RPS against a human opponent, but do not communicate with them directly - the interface is operated by the third side and only shows the choices (no chatting, etc). Your opponent is free to choose as he wishes. The operator intervenes with 50% chance (fair coin toss) by replacing your opponent's choice with rock (replacing rock with rock still counts as intervention). Of course the fact of intervention is not communicated to players.

Scoring rules: On "normal" rounds you and your opponent score against each other as in normal RPS. On "intervention" rounds you both score against the operator's budget. Your score depends on your choice vs. operator's rock. Your opponent's score depends on his original choice vs. your choice (so, yes, in this game you can both win/lose simultanously).

1. What is your best strategy in this game?
2. More interestingly: is the best strategy different if your opponent does not know about the existence of interventions (they change nothing visible on his side after all)? If you could tell him about it before the start of the game, would you?

That's my main question, but for the really bored, some more options. Consider a symmetric variant, where your choices as seen by your opponent also get replaced with rock randomly. Does the strategy depend on the operator's algorithm: option 1 - single throw decides no replacement vs. replacement in both directions, option 2 - separate throws for each direction, option 3 - always replace, single throw determines direction?

And what if your choices get replaced with something different than your opponents'? He 50% rock vs. you 50% paper, or he 50% rock vs. you 50% scissors?

And the most difficult one: review all these variants under "no ties" rule - in case of tie you replay the round, but the coin is not thrown again - if there was an intervention it will happen again, if there wasn't, it won't. This is a difficult variant - the same game might be a tie-resolving round for one player and a start of a normal round for the other, so you're never really sure whether your next round will get a throw. You will no longer see 50% rocks, the actual proportion depends on strategies of both players.

Man, coming up with such variations is fun. And details sometimes matter. Adding a side channel for communication between players also might change everything, as you can then cooperate against the operator... And from the player's point of view all these variants technically still fit within the bounds of the same simplified description: random choice of 50% rock, 50% human RPS player...

100 years? No, that's not how these problems scale. If it isn't P, then a macroscopic object (like the cat) is beyond our ability to simulate using deterministic Turing machines, period. The universe will not be habitable that long. Think about how many particles are involved, then realize that the algorithm scales exponentially. Even if our computers get a million times faster someday - so what, that's still not nearly enough. So, either we find a way to build fast and scalable nondeterministic Turing machines (we're nowhere near that goal) or the model is useless for macroscopic objects.

Yes, compression doesn't seem to be the likely reason, I'd look for other ones. But I'd be careful about using any publicly available comparisons of compression methods in this case, because it does not matter which one is better on average. The scientists know what to expect, how the image is likely to look, etc. They'd choose the best algorithm for this specific use case, and that might very well be a different one than for FB selfies... Not that I expect a huge difference anyway.

In other words it's time to start publishing your photos through your own account and others with false identification. Use different names. Make sure to reuse them at least often enough to make several options likely - if you use a different one every time, your own name on several photos will be enough to identify you. Use existing names, best ones would be of other people using this technique.

Build a large enough group of real people sharing their names and an even bigger pool of fake identities, automate selection of name to tag the photo. Let them sort THAT out.

Of course this is only good enough for screwing with their business data mining. For "targeted attacks", such as NSA trying to identify a person from an ananymous photo it doesn't help - they have different sources, you can't poison them all. Plus, if they can narrow it down to a short list of possible names, you're no longer facing this algorithm. After crosschecking with other, non-visual databases the most likely options will be reviewed by a living person. No, this approach is too weak to hide effectively.

Hell no. There's no comparison between Vista and 8 flops beyond the fact of being a flop.

Vista was an unbelievably badly executed step in the right direction. Bugs, half-baked ideas, better security done wrong, etc, etc... but essentially simply a next version after XP. With sufficient patches it becomes entirely usable and Windows 7 is essentially what Vista should have been. Failed version, not a wierd experiment.

8 is a completely different case. Total redesign of the interface, new APIs, new business model, etc. Very bad ideas, but very well executed. The system itself is quite good, fast, stable and well made, certainly not worse than 7. It is simply the wrong design, a forced change of the ecosystem. That's something you can't fix with technical patches. You either accept the fact and scrap the (huge and costly) project, releasing Windows 9 as Windows 7++ (or heavily patching 8 reverting it to old design, but that's a HUGE patch, essentially making it a different product), or clench your teeth and press forward, hoping that after some "tuning" of the UI users will follow you whether they want it on not.

Two completely different types of failures. Vista was the new Windows ME(*), Win 8 is the new Microsoft Bob (**).

(*) Except ME wasn't fixed by the next version like in Vista/7 case, because MS already had something new almost ready - having worked on the NT family in the server/workstation segment for years they were ready to deploy it on desktops (as XP), so fixing ME would be a waste of time.

(**) Except Bob was just another product, not a way towards a new sustainable business model for the company (the store)...

There's a polish idiom "pi razy oko", meaning a very rough estimate. Literally it means pi times eye (heh, pi day reference...). I find that literal meaning extremely useful for rough estimation of tasks, when not enough data is present to estimate better and/or the estimate is needed immediately.

Split the task into subtasks if possible, you'll get a better estimate. If you can't, don't - the estimate is very, very rough anyway.

Now apply the eye - using your experience give a pessimistic (important!) estimate of each subtask. Add up (allowing for parallel execution if you can allocate things to different persons). Now it's time for the "pi". Multiply the result by pi. No need to get precise, just 3 and some. What you get is the optimistic (!) estimate of time you need.

So far this failed me very rarely and not by much. The result always seems overblown to me, yet it's almost always a correct lower bound. Great tool to fight my own intuition where it tends to fail.

So.. yes, it will take longer than it seems. At least three times longer, so experience tells me.

Same here, and it works, with a large "but".

I got most of the obligatory and recommended vaccines as a child. For some reason though... not measles. Perhaps it wasn't obligatory at the time and I missed it somehow, maybe there were other reasons... I'd have to ask my mother.

Anyway, I never had measles and wasn't vaccinated. I'm perfectly healthy now. I don't have children, but most of my colleagues do, so for the past few years I had contact with people who could be carriers several times. For a grown man this virus is a much more serious problem than for a child. So, let's get vaccinated!

Nope. I lost enough energy to stop trying. One doctor laughed at me. One wasn't sure if it is at all possible to vaccinate an adult. Others were more competent but still - no. You can't just buy it, not without prescription and who will give it to you? I don't get it. Are they afraid of liability if something goes wrong? Is it the result of the public funding limits on different procedures? Maybe that, but I tried it at a private clinic, paying for the visit. Still a no: "you don't need it, don't worry".

WTF?!?

...which hasn't stopped anyone from using it - rationality is for the weak. We're wired for "eureka" moments - the curve fits so well, it MUST be right!

OTOH, technical analysis is also not a very good model of this, because economy is not a good model of anything in the real world due to an exceptionally strong positive feedback loop between the model and the modeled. A successful technical analysis "method" (meaning it worked for someone, that's statistically probable no matter how stupid the method is) may become actively used by enough investors to change the actual process, becoming a self-fulfilling prophecy. A given shape may have no meaning as such, but if enough people believe it signifies a change of trend... their reaction may cause the trend to appear, at least in short term.

In case of flu the feedback, if it exists at all, will be negative - overestimation may increase popularity of the vaccine, resulting in reduced number of cases. The reverse is also true, low estimates decrese the willingness to vaccinate. That of course completely ignores the question of how effective flu vaccine really is - we have to predict which strain will dominate; a miss would weaken the feedback.

Besides, why "on the Internet"? The assumption here is that it's somehow hard for the NSA to infiltrate an intranet - hard to believe given the wide choice of tools they have. And unencrypted MySQL on the intranet is common.

Heh... Easy question, unfortunately.

Top priority: prepare an easy and painless way out. Guns are illegal here, so it would take a bit of thinking, organizing, saving money, etc. Probably the best solution would be an international trip to a clinic that will help me, but I would need a backup plan if someone decided to stop me. Better do it early and be ready for later, with a plan simple enough to execute when the illness already has a significant effect (but before it makes me forget I have that option). Later I may not be able to do this and noone will help me. Hell, I wouldn't even ask for it, I don't want that helpful person to go to jail.

Oh, I could have other priorities, if I could achieve this just by making my wish clear. But as long as euthanasia is not legal here, I'd have to rely on myself, so waiting too long would be risky. I will not reach the final, infant-like stages if I can help it. I prefer to keep my dignity, thank you.

So, this is the most important thing. Number two is obvious too - research into current best practices and applying them (diet, activity, training, whatever). Even if it buys me just a few more months of mostly normal life, it's worth it.

Not to suggest that anyone should do the same. If your views or priorities are different, feel free to do whatever you want.

Oh, the naivety... The winners are just less clearly defined now. Unless it comes to actual combat, that clears things up. Let's hope that doesn't happen.

Anyway, it seems like history will repeat again. Just like 1938. Diplomatic pressure, discussions, etc. right until the West is under attack. Oh, wow, how could that happen? Avoid war at all costs, sure, but find a good way to stop the conflict or don't be surprised later.

Oh, and focus on the facts. This is not the time to discuss whether the political shift in Ukraine was legitimate or not. Focus on the actual territorial claim. Right now Russia is trying to tear Crimea apart from the Ukraine by military means. Focus on this fact. Maybe Crimea should be part of Ukraine, maybe not, doesn't matter at all. The use of the military is the problem. Or just ignore it, fine. Just don't be surprised if it doesn't stop there.

And why would we want it? Seriously, we have good coffee here.

I keep hearing things like "only a douchebag would take the last of the coffee without making a new pot" - seriously? You drink random drip coffee from a pot that's been heated for the last hour? I would, if I had no choice and needed some coffeine, but... come on!

My espresso machine at home is 4 years old, cheap (no auto) but good. Coffee from a nearby coffee shop, their own mix, roasted in house every second day, pure arabica of course. I grind it at home, enough for two days (yeah, I know, too much too rarely, but I'm not that much of a perfectionist). A great espresso every morning. Nowhere near as good as what I tried in Italy, but way above average here.

Let's take the cost of the machine and the one service cost when it broke last year. The cost of the grinder (a good one actually). The cost of coffee per 100g and serving size... 4 years... 4 cups daily on average (low estimate - I don't live alone)... Something like $0.43 per serving, excluding cost of electricity and water (cheap but low calcium, in large bottles, no real difference in taste from tap water, but good for the machine, doesn't require decalcification as often - my time is not free). Even with those, still under $0.5. Give me one reason to buy this.

Even at work if I need coffee it's the same coffee as at home, always fresh, made as needed, but from a drip machine. Not nearly as good, but oh, well... Two years ago we had espresso machines at work, but they were too cheap for heavy use - they slowly deteriorated and after too many problems thay were eliminated. I'm still hoping they will buy a couple of good ones. Another team bought one for themselves (legal if it stays in their kitchen), so maybe we should just do the same - except we don't have a separate kitchen, so it would have to be good enough to handle about 40 espressos per day - not the cheapest model... Maybe we could get the neighbouring teams to chip in...

I'd never buy any machine that would tie me to a single brand of coffee, because it simply is not good enough. I've searched for a good source for years and I will stick with it until I find something better. Why would I buy something that sat on the shelf, roasted and ground, for hell knows how long?