The developers building vision algorithms don't typically create their own datasets. They purchase archives of images, and a lot of these problems stem from how many samples of each type are in those archives. The Google team likely has a giant database of human faces that it works with, and the ethnic frequencies are probably either the result of choices made by whatever origanization compiled it (and for whatever reason they compiled it) or the ethnic breakdown of the userbase of some app they used to grab the data. It's extremely unlikely that either of those will produce the same number of samples of every ethnic type.

It's also one thing if this was a program just designed to distinguish between different people. But it looks like it's trying to recognize objects of all sorts and distinguish between people and just about everything else. That's a hard problem, and the only response to this sort o thing is to take a regular failure case and feed it back into the training data so you can hit the next regular failure case. Hopefully it will be less coincidentally embarrassing, but it will definitely be there. Perhaps confusing bald men with balloons or something like that.

But I also think people underestimate how much skin color affects machine vision problems. I spent years in the biometrics industry and one consistent fact is that people with darker skin just don't provide as much easy-to-recognize detail as people with lighter skin. There will be more misclassifications as long as the image is taken using the visible spectrum. To a computer extracting features, dark skinned people and gorillas are both human-ish face shapes with a particular color range and somewhat indistinct geometry due to weak contrast and shadows. Distingushing between those two sets just isn't as easy as distinguishing between fair-skinned blondes and gorillas. You can make that decision just by looking at the color histograms and not even bothering with geometry.

The classic Schneier Applied Cryptography is a great read for anybody who wants a good starting point on the basic concepts and practical considerations. It's technical-ish but conceptual rather than mathematical and leans toward describing what the various crypto pieces do, why they exist, and what they're used for. To get a good intro to some math, try The Handbook of Applied Cryptography. If you have a little bit of number theory and are willing to do some exercises up front, the book is largely self-contained and very well written. It's free for personal use, but nobody I know regretted buying a hard copy.

White dudes had the right to marry white women and black dudes had the right to marry black women before Loving v. Virginia, but the court still ruled that prohibiting white dudes from marrying black women and black dudes from marrying white women was an equal rights issue. Now it seems like a no brainer, but at the time the same logical argument was made. I'm guessing we'll have the same perspective on it in another 20 years.

Property, power of attorney and inheritance are pretty much it as far as I'm aware. But that ain't nothing. What happens to your property when you die, who can make decisions on your behalf if you're incapacitated, who has a valid right to raise your children, etc. are all pretty important issues. Having everybody more or less agree on how it's done through marriage and family lines was pretty convenient. If we do away with it, we'll have a lot of issues to work out, since the government ends up in the middle of all of those disputes once they go to court.

Gay marriage fits pretty neatly into those paradigms, so it seems like a no brainer. I'd be all for legalizing polygamy as well, but the 11 algorithms that are assumed in the law don't necessarily scale to 1N or NN. It seems like it would be worth coming up wtih some more baseline principles that would allow us to define who is "family" in a general way.

Right, but they have the option not to now. Instead of having to tell us what they'd do to smack the gays around, they can signal their cultural affiliation by tut-tutting the ruling and saying, "This is terrible and I agree with you, but my hands are tied." They can make as much or as little noise as they want and then as the electoral tables turn, they can taper off. Before the Supreme Court ruled on it, it was a very real issue for the legislature and the executive. There was going to be a critical mass that demanded gay marriage at some point, and the last people on board were going to have to go on record voting against it. Better to be able to blame the other guys and not be on record doing anything one way or another when you're on the wrong side of history.

how does life, liberty or property equal marriage? not just gay marriage. Marriage for anyone?

"...nor deny to any person within its jurisdiction the equal protection of the laws." You can't just arbitrarily treat a group of people differently under the law without a good reason to justify it. That's why you can't make a law that says that only white people get free public education but you can make a law that says that only poor people get food stamps. Free public education and food stamps may not be fundamental rights in the traditional sense, but the simple rule is if we're handing them out, we have to hand them out equally to everybody unless there's a very good reason not to.

Lastly, the words of the constitution do not grant unlimited flexibility. If we want to change the 14th amendment or any amendment to the Constitution then the people need to step up and work with their own states and legislatures (NOT THE GOVERNMENT) to do this. This is what the amendment process is for.

Point blank question: Was Loving v. Virgina wrongly decided?

What's to stop three people from wanting to marry? I don't mean to be a conspirator but according to the language that I see there is nothing that can stop it.

What's missing from this is a reason why we should be concerned that it doesn't "stop" three people from marrying. "Look! Three people getting married! We need to do something about this!" The main issue I can think of it is that it's structurally tought do with the way some of our laws are assumed to work. Other than that, meh.

I am waiting now for the first lawsuit to appear about a pastor at a church won't marry Jane and Sally because of the pastors firmly held beliefs and the core doctrine and tenants of the church's faith.

And you'll watch it get tossed out on its ass the moment it's brought. Churches have always had the right to decide the rules for their rituals. Interfaith marriage has been a right for a very long time, but nobody has yet forced a church to marry a couple in contravention of the religious rules of that church.

This ruling solves a very real problem for a lot of people and the problems people claim it cause are way out at the margins and frankly unrealistic for the most part.

Or re-interpreting a part of the constitution in a manner that would have mortified the people who actually wrote it?

People write laws that are interpreted and used in ways they don't expect all the time. That doesn't make them any less the law. If you write something that basically says, "The government will treat people equally," and privately assume that the people you're not treating equally will probably continue to be treated unequally, you'll probably be surprised to find that the rules you actually wrote down outlast the social norms of your time.

The first easy step would be to announce that the mortgage interest tax deduction will go away, reduced by 5% every year for the next 20 years. That's one that creates bizarre incentives and is basically just a transfer from taxpayers to banks. Tightening up the rules for the GSEs would definitely help. One idea that I particularly liked is to ensure that anybody selling a loan to the GSEs would have to keep a small percentage of it on its own books.

And before someone swoops in and says "well then all those currently renting and unable to buy will go homeless!": what do you think the people owning the rental properties are going to do with a bunch of excess property that's no longer of any use to them when they can't rent it out for profit? The only way they can benefit from it then is to sell it.

That would definitely drop the price of housing down to a point where more people could afford it. But I think you're missing a few nasty side effects.

1) There are a alot of people with no money. Net worth of zero or less. They have enough income that they could pay rent, but they literally don't have the cash to buy and nobody will loan them money.
2) A lot of people prefer to rent for a number of perfectly valid reasons.
3) The ability to build a building and rent the units in it for a certain price creates an incentive to create the building in the first place. With no rent option, that incentive is reduced. The plan to squeeze the properties out of the hands of the rich only works once the properties are built and in the hands of the rich. Once you've occupied all those, you'll want somebody to build more of them.

The step most people mess when they do this exercise is to weight the various goods by how much of your income you actually spend on them. When most people estimate inflation, they seem to assume that all they do all day is eat bread soaked in gasoline. People buy a really large variety of things over the course of a year (including, say, the fraction of the refrigerator that wears out during that year), and taking a sampling of a handful of goods to build a cost of living index usually produces unreliable results.

Developers are building new apartments as fast as they can--luxury apartments that charge higher than market rates, further inflating the market.

The additional luxury apartments create downward pressure on prices, not upward pressure. It's the demand for apartments in general that drives up the prices. If they weren't building the luxury apartments, the people who wanted those luxury apartments would likely just outbid less rich people for less luxurious apartments.

There are actually places that license pyschics. I have no idea what would have to happen to lose your license to practice.

One thing I tell everybody who fills out the SF86 is, KEEP A COPY OF YOUR SF86. You'll probably fill it out more than once in your career. Starting from scratch is a gigantic pain and errors creep in if you have to look up older stuff from primary sources.

Actually - yes I am. I watch people avoiding the police. I watch people saying "Yes sir" and "No sir" to the police. I watch people groveling in front of the police.

If there's one thing the data has shown us, it's that we're all much safer in an interaction with the police if they honestly believe we might shoot them.

I address police in one way, and one way only. I address them as equals. I am a free man. Cop says "Stop!" I say, "What for?"

Do you reach into your waistband, just to make sure they know that you're not going to take any crap?