579 karma · joined September 1, 2012
Are you sure? The point of the article is that even if humans won't willingly start a nuclear war, they are so used to the idea of nuclear weapons just sitting there, that they stopped fearing them and become careless. This may one day lead them to accidentally give, or to allow their automated systems to give, the launch command.
Also, if your application has fairly stable load, then startup times are irrelevant 99% of the time, since you can just run persistent servers.
> The company partnered with Wrapify, a San Diego startup that pays drivers to cover their cars with branding and data-tracking technology. Each car is equipped with sensors that feed live analytics to clients so they can see where their ads appear on the road, and how many people can see them. (Wrapify uses telemetry data from the car to track drivers' real-time location on a map, and pairs it with anonymized data from other sources to understand how many cars nearby see the message.)
Am I the only one who finds this creepy?
Compare with horizontal scaling, which in this case translates to large numbers of individually-capable units acting in coordination.
It's just too damn hard and expensive trying to protect a single valuable target against an ever-increasing number of threats.
The good thing is that "runtime" doesn't have to be production - you can experiment with live data in a REPL and flush out most of the bugs, which of course doesn't diminish the need for good tests.
I wouldn't recommend it for teams consisting only of junior developers and/or large codebases, where it can be hard to maintain good hygiene in the long run.
The first project used Spark, so Scala was a natural choice. One big-ish issue at the time was that our team didn't have anyone with experience in the language, so the code base ended up being very Wild West-like, with different styles mixed and heavy usage of implicits. However, none of my teammates had much difficulty becoming productive in Scala (not being an expert, but just being able to produce sane code, quickly), even though they came from mostly non-functional backgrounds (Java, C#). I did struggle with the type system initially - needed some time before it all "clicked", e.g. when to use 'flatMap' instead of 'map', things like that.
The second project uses Akka, which is by itself pretty neat. This time however, I'm the only engineer with experience in Scala and writing production code in general. I feel like the type system provides a lot of protection against silly mistakes that newbies sometimes make (especially with the poor test coverage, such as it is), and the code is much more concise that Java would've been, so yay smaller code reviews! The rest of the team is familiar with traditional C-like syntax, so it's easy for them to get started. I try not to push too hard for the functional features initially, instead showing people how their imperative code could be rewritten in a more clear and concise way.
One major problem is the IDE support. It's 2016, yet IntelliJ is still slow as hell with Scala, and I've abandoned Eclipse long ago.
I do notice that the language seems to promote in some people a certain tendency to write esoteric code and use some very cryptic features just for the sake of it, but I haven't been bitten by this in the libraries we use, and try to be pragmatic in our own codebase.
Lastly, the value of the infrastructure - the JVM, the rich set of libraries for Scala and Java, Scala itself with stuff like Scala.js and Native - should not be underestimated by anyone who considers the language for production use.
Oh, and I do have some comments about Clojure as well, if anyone's interested.
Some issues though:
* battery only lasts around 90 minutes on full settings, and half of that if you're compiling something heavy. This is a used laptop though, so no idea how many battery cycles it had when I got it.
* switching to an external monitor is annoying. Out of the box it doesn't remember the display layout settings (mirror/extend/etc). If I switch off the on-board GPU and use the discrete one, the VGA port is unavailable. Sending sound over HDMI doesn't always work.
* wifi sometimes gets into a weird state where it can't find any networks. Resetting the driver helps though.
* the only available AC adapter weighs a couple of pounds and is huge. Haven't had any luck finding a more reasonably-sized aftermarket one.
* the speakers are too weak even on 100% volume.
> First, Mint has bank-level data security. That means we have the same level of encryption your bank does, along with outside third-party verification through Verisign and Hackersafe. We also have routine security audits where so-called “white knight hackers” try to break into our system — they’ve never been successful. We also have bank-level physical security. Our servers are located in an unmarked secure building which requires a palm scan to gain entry. After making it past guards, you have to go through a “man-trap” where one door will not open until the other closes and you again have biometric access. Once you get inside, our servers are in a locked cage monitored with 24/7 video surveillance. Get inside, and the racks themselves are locked. Break those open, and our hard drives are encrypted. It’s seven layers of protection. All that’s missing are the electrified floors…
All of this is provided as justification that 2FA with banks is not supported (original topic of the linked thread).
> Ammon Bundy (I did not meet him during my visit to the refuge) may or may not know the history of land use in the West, but there will be no repeating the free-grazing era of the late 19th century. Not in the fastest-growing developed nation on earth, on a planet that will soon play host to nine or 10 billion human beings. Nothing will be free. What the militants are asking for is almost exactly what more mainstream political leaders like Utah state Rep. Rob Bishop or the American Lands Council, now headed by Montana state Sen. Jennifer Fielder, say they want, too. The Malheur occupation, with the incessant press coverage in its early weeks, was the soapbox for disseminating payloads of misinformation about America’s public lands, about their management, about how and why we have them. Every soundbite was delivered to further the goal of privatization.
> In 2012, when the average wedding cost was $27,427, the median was $18,086. In 2011, when the average was $27,021, the median was $16,886. In Manhattan, where the widely reported average is $76,687, the median is $55,104. And in Alaska, where the average is $15,504, the median is a mere $8,440.
Luckily, these are not national figures, merely numbers reported by visitors of a particular website. Unfortunately, I couldn't find any more representative numbers, especially with expenses measured against income level.
[1] http://www.slate.com/articles/life/weddings/2013/06/average_...
> Avy added there will be a few other stipulations in the contract, like a clause that makes one person responsible for the entire debt if a marriage ends because of abuse.
[1] http://www.geekwire.com/2015/this-startup-will-pay-for-your-...
> However, keep in mind that such potential dangers are quite far from us right now. There has been a tendency in the press to blow this up out of proportion to attract attention. It could hurt AI research. Instead, we should encourage research towards better understanding these issues, to build an understanding that can, if the day comes that it becomes something more concrete to deal with, prepare us better and provide us with safer AI that respects human values.
> To match the increased power that science and technology give us, the law of the jungle is not appropriate: we need an increase in individual and collective wisdom, for our sake and that of our descendants.
Yoshua Bengio
> No, provided we stick to a simple rule: don't create AIs with goals of their own. AIs can come up with their own subgoals, but only in service of the goals we set them, and within the constraints we specify. This is how all AIs work today, and as long as they keep doing so, they can be infinitely intelligent without being a threat to us.
> Of course, human nature being what it is, sooner or later someone will try to create a self-seeking AI. To deal with that, we need what William Gibson called the "Turing police": good AIs that catch bad AIs in the same way that cops catch criminals.
Pedro Domingos
The description sounds awfully similar to Spark, yet they say it's "built from the ground up", which I assume means "from scratch".
Corollary: if you find that your job doesn't let you learn (e.g. your boss just wants you to churn out as many web apps in as little time as possible), leave immediately (or better, try talking to the boss first :) ).
However, as ef4 and others have pointed out, the real question for you might be "what are the problems I'd want to spend my life on?" Not all important problems are hard - some may even seem too unsexy to be worth working on, but the impact can be huge.