so it would be $0.03 per can rather than $0.01. Oh, darn.

Are we ushering in the age of Nuka-Cola?

It doesn't completely solve the problem, but it allows for some recycling where there is no solution today.

Pull burned carbon from air > make diesel > burn diesel > go to step 1.

How is this not preferable to:
Drill miles into the ground > inject god-knows-what fracking liquids > extract oil > flare off natural gas found with oil > ship oil to refinery > refine oil > burn diesel > go to step 1.

I don't understand the mentality of "It doesn't solve every aspect of every problem neatly and completely, so it's shit and shouldn't be pursued."

And you're forgetting that once someone uses the diesel, it's CO2 in the atmosphere again.

What, did you figure that Audi was just making diesel to pump back into the ground and forget about?

And then burning it again. We're not just making barrels of diesel to look at like art - it's going back into a cylinder to be compressed with air until it explodes and blows out the tailpipe again.

It may also mean that it was more secure than what the "government-trained information security specialists" would have done.

Isn't baseless speculation fun?

Oh noes, don't read my unclassified email that is likely to be public record after a FOIA request!

These are either the shittiest spies ever, or someone is trying to drum up nationalistic sentiment. I'll get worried when they're reading secure diplomatic cables that haven't already been posted to Wikileaks.

Yeah, because it's really hard to unplug the cable from that Atari console and plug it into the Apple II. It requires electricians with union cards and stuff.

If you use CloudFormation, you get auto-scaling. In fact, if you write your own CloudFormation template, you can get easy drop-down menus of minimum / maximum instances, and how many to scale at a time.

No, Amazon doesn't have direct publishing from Visual Studio, but we've worked around that by using a Jenkins continuous integration environment and Chef. Jenkins automatically builds any code merged into watched Git branches, and if the build succeeds, drops it into an S3 bucket. In 15 - n minutes, the dev / stage servers will run Chef, which sees the new version and deploys it per the app's recipe. It all happens as soon as someone code reviews and merges the branch back.

With RDS, you've got so many options for backup it's not even funny. You can have read replicas in different availability zones. You can have snapshots. You can use Cloud Protection Manager to automatically snapshot. You can use replication to pull data out of RDS to a database running on an EC2 instance for ETLs, or to ship it somewhere else. You can use that same replication to have it sit on an instance where you are running a database backup daemon of some type, which offloads that work from your production database. The only thing you don't get with RDS is access to the OS image it's running on, or the filesystem. Anything else is fair game.

As the other guy said, Azure holds your hand more, and is much more Microsoft-centric. But the offerings are not nearly as deep, or as wide. Especially in networking - with AWS I create multiple private clouds that are walled off from each other and geographically separated, and only allow the traffic that I want across a VPN tunnel for DR purposes. I can peer together AWS virtual private clouds to allow network access between the all the instances, but retain security for departments to only manage instances in their account, in a web form that takes two minutes to set up. Also, because of some luck, we're leasing space in a datacenter where Amazon has a DirectConnect node, so we can physically peer our company's network with a few VLAN tags, a fiber pair, and setting up some BGP neighbors; which also saves money on the network data because it's no longer going out through the Internet to get back to our datacenter.

And they absolutely don't make up for it when someone pays their Prime membership and only orders 4 things a year.

In addition to EC2, there are lots of other services that are encompassed under AWS that compliment EC2 nicely - RDS is their service for standing up easy database instances that take care of most of the configuration headache associated with the big relational databases out there (pgsql, mysql, mssql, oracle). Route53 is a scriptable DNS service. CloudFormation gives you tools to automate standing up entire application stacks including DNS records, load balancers, application servers spread across availability zones for redundancy, etc. OpsWorks gives you a Chef-esque service for managing software deployments. IAM roles allow you to grant servers access to other AWS services without having to deal with certificates / passwords / keys in an atomic fashion. There are numerous other services that I'm not even using right now, but exist as replacements for things we've already wired up in EC2 previously to Amazon announcing them.

It's a hell of a package that makes things like Microsoft Azure look like a joke in comparison, and is cost competitive with building out your own datacenter as long as you use it properly, and think about what you're doing a little - make sure you back your shit up out of AWS so that you always have an "off-site" copy, for example. And if you use something like Chef / OpsWorks, you can recover from a disaster practically anywhere if you have your cookbooks, source code, and data backed up.

Who gives a shit?

How about "Civilization." It's hard to think of other modern projects that have advanced the knowledge base of humankind as far as the Hubble Space Telescope. Not bad for being a "multi-billion dollar flop" when launched.

Of all the things NASA has accomplished, this is one of the big ones.

Following your "logic", Best Buy is responsible for the millions of computers that get infected with shit from running copies of Windows that were purchased at Best Buy and not patched / maintained? Because Best Buy just "pays for the brains of these app developers and then they resell it" ?

Brilliant.

In no way does what the guy is describing magically allow code to take control of the full OS. If an application is executing, and then executes a maliciously crafted dylib, that dylib is still running as the user who executed the parent application - a.k.a. not root unless you've bent over backwards to re-enable the root user and log in as root because you completely hate security and best practices. If it wants to do something outside the permissions envelope of that user, it will still have to ask permission just like anything else on the OS; even if you are running as admin - all that gets you is the ability to put in your password to allow it, rather than have to click cancel. The only way around that is to also combine a privilege elevation exploit - and now we're getting into the incredibly improbable that you could find a signed app that would do both without a user seeing anything odd.

At the end of the day, GateKeeper wasn't designed to prevent that anyway, and this guy is presenting a massive straw man. GateKeeper was designed to give you a decision point between clicking on the random thing that appeared in your downloads folder, and getting owned. That's it.

If it works for them, and they are happy with it, what the fuck do you care?

I really don't understand this attitude that everyone must use the same tool for all jobs, ever.