599 karma · joined September 13, 2010
[ my public key: https://keybase.io/zattack; my proof: https://keybase.io/zattack/sigs/ju1icv42KBBcbTu8e4RQoq4SvAJcs3TODCRCzfrL1Nw ]
This especially doesn't work against multiple opponents.
"Doing X in PHP is a bad idea"
"Well it's only a bad idea if done by unskilled developers"
But about machine learning... a system can learn from a set of outcomes which are the result of a "biased" system, that is to say tainted with an incorrect Bayesian prior which is not properly corrected, such as let's say courts in a racist society or whatever. It would learn the same bias. Because its goal is to maximize compatibility with the outcomes that the humans did. So it perpetuates those weights.
The problem is that we don't know whether the human decisions matched the reality. "Did the person commit the crime" for example. We might have to wait until more unbiased estimators for such activity come along, and throw away old historical data.
It's sort of like when Black-Scholes became a self-fulfilling prophecy for valuing derivatives, but only after it became widespread. The market started using Black-Scholes to value derivatives, so it became the best model to predict the value of derivatives. But until then, other models might have fit the historical data better.
If people could rank their preferences then people like Bernie and Bloomberg would actually run without fear that they might "take away votes from Hillary" and therefore help Trump. In fact, the two-party syste would give way to something better. Only the party elites would NOT want this... but after this election, even they probably do.
2006: http://english.ohmynews.com/articleview/article_view.asp?men...
2010: http://www.mediapost.com/publications/article/127409/faceboo...
2013: http://www.pcworld.com/article/2047804/facebooks-new-face-re...
2016: http://www.nytimes.com/2016/03/03/business/international/fac...
http://qbix.com/platform/features
Would like to get some feedback on it, from anyone who looks through the features.
If you're going to argue that we as a country of laws should deport them all back, then I hope you and your family never smoked pot because you broke a law. And since we are a nation of laws - including minimum sentencing laws which the prison industrial complex loves - how would you like it if they looked for you and put you in jail for a victimless crime? Drop your double standard. The Mexican immigrant is better than the potsmoker because they fled violence, wanted to make a better life for their family AND helped do the jobs no one else would. The potsmoker chose to smoke and helped no one except the drug dealers.
Plus we did that already, and it was a disaster: https://en.wikipedia.org/wiki/Mexican_Repatriation -- an estimated 1.2 million US citizens were deported. Plus until 1965, immigration was unrestricted from Mexico and Canada so many of the 11 million broke a law by staying, but not by coming.
You have heard all these myths. The fact is, immigrants have higher labor participation, lower crime rate than the native born population. Especially the illegal immigrants who are afraid of being caught by police and deported. Illegal immigrants do NOT get money from the federal government - if your city pays them take it up with your city. But they pay taxes like everyone else, including sales tax and property taxes. So they pay into the system and get nothing back. You want to deport them all and break up their families so you will end up picking crops, and think this is the way to bring jobs to USA?
https://en.wikipedia.org/wiki/Companies_of_the_United_States...
It should be the other way. There should be a law saying that a company must meet some hurdle to store personal information for a long period of time. At least disclose it publicly. It's so ironic that the EU cares about websites disclosing that they store cookies, but not that they store conversations indefinitely.
The law would put a chilling effect on storing personal information indefinitely. But, like carbon emissions, it will only slow the progress towards a future where all the carbon is released from the ground / all the data is stored and analyzed by AI years from now.
How about greenhouse gases in the atmosphere?
On the other hand, I'm not sure how to explain the presence of homosexuality in a population when the theory of evolution predicts that any such trait (which leads to much lower genetic fitness) would have disappeared a long time ago. How does it stay at this level from generation to generation?
Didn't I read just the opposite in an article on Hacker News just a few hours ago? It claimed that researchers were mistaken in the past, yet the studies cited here are from the past 10 years.
http://www.slate.com/articles/health_and_science/science/201...
http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1071721/
More balanced:
http://journals.plos.org/plosone/article?id=10.1371/journal....
And this company is actually registered? Or it's all fake?
That is just security theater. It's worse than useless because it makes you think you're more secure, when you haven't prevented attacks are all.