And maybe more importantly, if it disagrees with the humans, can we figure out why? This is a problem with a lot of machine learning applications, but there aren't many where understanding the decision-making process is more vital than it is here.

He says. On the internet.

From RT:
http://rt.com/politics/177248-...

Such authors will now have to register with the state watchdog Roskomnadzor, disclose their real identity and follow the same rules as journalists working in conventional state-registered mass media.

  The restrictions include the demand to verify information before publishing it and abstain from releasing reports containing slander, hate speech, extremist calls or other banned information such as, for example, advice on suicide. Also, the law bans popular bloggers from using obscene language, drawing heavy criticism and mockery from the online crowd.

So.., now you're not legally allowed to lie to a large number of people or incite violence based on those lies. Gee. That's bad how? Might be nice to have something like that in the West, because right now it's perfectly legal for FOX News to outright lie to their viewers.

Russia, like any large nation the US hates, (see Venezuela) must defend against the standard CIA tactics used to de-stabilize governments and population bases through grass roots propaganda tactics. Forcing creeps and liars out of the game seems like a pretty good way to do this. You don't want to be forced out? Then follow the law and back up your claims with fact checking verification of what you are writing, don't use hate speech and don't incite violence. How hard is that?

There's a reason you're not allowed to yell "Fire" in a crowded theater, and this falls neatly beneath the same rubric.

Honestly, think of the gossips and cruel kids in school spreading lies in deliberate attempts to undermine healthy energies. Putin has the guts to whip the carpet out from under such types.

So now, once you reach 3000 readers, the Russian government says you are a news source with real pull and must start acting in a manner befitting such responsibility. Is 3000 the right magic number to have picked? I don't know, but it makes perfect sense to draw a line somewhere.

Of course, any law can be abused, but right now I don't see this as an abuse. I see it as a sensible measure as Russia is under increasing media attack by a truly psychopathic nation whose leadership is completely disconnected from objective reality, has a tail-spinning economy and seemingly bottomless war lust. Of course you have to take measures to protect your populace from that kind of sickness.

But naturally, this proactive move is being spun with wicked and/or childish glee in the West (depending on whether you are CIA or just ignorant and easily led).

They would get them back and then punish them and then separate them.

Exactly. If that's what he deserves, then truth will out.

And I have seen an awful lot of people saying that he wasn't worth any particular effort to get back, which is pretty close to "let him rot." That's just mind-boggling to me.

Heh. Yeah.

I was at Minot for five years, which seemed particularly like exile after having been in England, about an hour away from London, for two years before that. I will say that it wasn't quite as bad as I expected it to be when I got my orders.

Were you at Dover? I've always heard that's kind of the East Coast's equivalent of Minot. [1/2 g]

So in that sense this is the most elegant natural solution.

No, biological processes are inherently non-deterministic, and this becomes more apparent the smaller the scale. At the genetic level, it's all about probabilities. I suppose you could argue the same about computation since circuits are now getting small enough for quantum effects to show up, but I don't think most programmers are explicitly modeling random bit flips! On large scales, when you're talking about big programs with lots of different possible inputs, it's often more effective to model them statistically, I agree, but the underlying processes are still quite different.

Co-evolution only looks "co" on very large timescales; every new trick our immune systems have come up with has been in response to something a pathogen already came up with. Sure, there always can (and will) be new plagues, whether the victims are trees or people. I just think they're a whole lot more likely to come from the nigh-uncountable number of random "experiments" taking place in the wild than they are from anything done in a lab.

That last bit would be a lot more persuasive if it weren't for the concept of limited liability. The whole idea of corporations owning assets, signing contracts, etc. is that the owners of the corporation are to some degree insulated if the corporation "does" something such as breaking a contract that could lead to the loss of those assets--but it really ought to work both ways. As things stand right now, the privilege pretty much only seems to go one way.

Honestly, I think that fear is overblown. Vertebrate pathogens have had hundreds of millions of years of optimization in the most ruthlessly selective "laboratory" ever known, and while there are obviously some pretty deadly ones out there they haven't managed to wipe us out yet. Nothing we do in a lab is likely to come close, in terms of coming up with something that can spread wildly on its own.

I used to work between a synthetic bio lab at one end of the hall and an infectious disease lab at the other. Ask which one scared me more.

In a lot of ways, it is similar, but there are some important differences. The biggest one, I think, is that programs are (or had better be!) deterministic: make a particular change and a particular thing will happen every time. Living systems, even relatively simple ones like yeast cells, are stochastic: make a particular change and the probability of a particular thing happening increases or decreases. What you're counting on when growing a culture of mutated cells is that enough of the cells will behave in the desired fashion to make the behavior of the colony predictable, but the underlying randomness remains.

And up until now, we didn't know if those right conditions ever occurred for asteroids. Now we do. That certainly counts as a discovery.