Want to read Slashdot from your mobile device? Point it at m.slashdot.org and keep reading!


Forgot your password?
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×

Comment Re:Important milestone (Score 1) 115

There was a story on Slashdot a while ago about the world's largest hedge fund replacing their fund managers with computers. That's not really that impressive though, since many studies have shown you can replace fund managers with monkeys flipping coins and get the same performance.

Many of the best learning systems are currently taught in an unsupervised way. They're fed stimuli and form their own internal model. Finally they're given a minimum of supervised training. Like a baby gazing around at the world for a few years then being told that the fluffy four legged thing is called a cat.

Comment True. Challenging assumptions, bending rules (Score 1) 21

I think there is some truth to that. I wouldn't do well in the Army. My natural tendency is to challenge assumptions and manipulate, if not break, the rules. This personality has served me well in my infosec career.

My tendency to always think about what I can get away with fits infosec well, but probably not DoD. It has also meant that I have to be very careful about ethical and moral behavior. Since I'm always thinking about how I *could* steal something or how I *could* spy on someone, it would be easy to start actually stealing and spying of I'm not on guard.

Comment but wait; there are markings (Score 1) 64

The abos are not so innocent as the liberals want to portray them after all.

Here's the thing: the upside of inventing a writing system is world domination; the downside is finally having to admit in public that you are a real ass (and always have been).

In the above, "you" is a set of nesting dolls, innermost being the fifty-year-old white male technocrats of western European origin who treat Wikipedia as their private, personal playgrounds (thence to aging white European males, white European males, white males, whites altogether, etc.)

Here's the second thing: after a society invents writing, soon the society has written myths (with serious legacy entrenchment) that innocence preceded the current sad state of affairs (how-far-I-have-fallen porn, not that the larger consequences can't be remedied by kneeling under the right cumulus cloud for a thoroughly abject sixty seconds).

Society will re-invent writing over and over again (movable type, Movable Type) before the reversal of true illumination makes the least headway: that the human asshole apogee was attained circa the advent of the original edged weapon.

As far as the abos go, they all need to repeat to themselves "there but for the grace of God go I", unless they think their ancestors truly enlightened enough to not have had even the most remote possibility of inventing any form of written record, whatsoever (best if you're not much past the wreathie leafy loin cloth, because any loose thread threatens to quipu a long record, and then immediately you're on the outie asshole train along with every other post-prehistoric posse of mugs, pugs, and thugs).

Comment Re:Firefox is back! And windows exploit more $$$? (Score 1) 46

Windows kernel exploits are worth more because they're worth more on the open market (because that's where the corporate data is and corporations pay ransoms). pwn2own has to compete with the black market, after all.

Wrong. All of these prizes are far below what a zero-day exploit is worth on the black market. This contest is not a way to overbid the black market; rather it is a way for white-hats to showcase their skills and bring attention to vulnerabilities.

The prizes a set to reflect the expected difficulty; the hardest target - the ones that involves the most work - pays most. Virtual machine escapes are considered really hard because of the very limited attack surface.

Windows 10 is considerably harder to crack than Linux and OS/X. The latter 2 still have *far* to many services running as root and still exposes a lot of SUID root executables. Windows 10 has also adopted many of the EMET anti-exploit techniques. You'd have to harden Linux with grsecurity to achieve the same level.

Comment The dam is valuable, the parking lot crack not muc (Score 2) 242

> Your thought process is akin to saying it makes no sense to spend $5k to patch a 2" crack in a dam because the crack is only 2".

No, the dam is extremely high value, therefore you pay attention to it. When the Banqiao hydroelectric dam failed, it killed hundreds of thousands of people. So the dam is at the top of your "most protected" list. What I'm saying is this:
There's a 2 inch crack in the dam, and a 2 inch crack in the parking lot. What's your first step? Your second step?

Obviously your first step is "fix the crack in the *dam*". The correct second step is less obvious - look for more cracks in the dam. You shouldn't worry about the 2" parking lot crack until you've double checked everything about the dam. Again, see Banqiao.

Comment Doesn't hurt, besides performance and trust (Score 1) 242

You certainly can do both. There will be a performance hit, small or large depending on cipher mode. You should double-test your backups in case either layer of encryption fails. I would recommend using a fast mode for the full-disk, keeping in mind it won't be NSA secure. So thinking about privacy, you'd pretend the full-disk isn't there - it's just a backup just in case.

Comment One of us is misunderstanding the other (Score 2) 242

FYI I've been a fulltime security professional for 20 years. My advice is based on what I actually do when your bank hires me to test their security, how I can actually hack your accounts.

> No, the problem is, you try to seperate, what seems important and confidential to you. And there is the mistake.
> Because it requires you to think about what's confidential all the time. ...

> reading some private e-mails won't hurt now, because if they are left in the cache in your firefox profile

I never said "encrypt one file at a time". I said encrypt YOUR files separate from your (soon to be ex-) wife's files. That includes /home/allo/.cache/mozilla/firefox/

Obviously you might *also* separately encrypt your most important files, such as a password manager datastore, a second time. But no you don't have to think about what to encrypt, all of your personal files are encrypted, including your browser cache.

> Why would you encrypt /home and not /? Is there any reason preventing / encryption? No. ...
> So you install your system, make a checkmark at "full encryption"

That SEEMS like a good idea, if your understanding of encryption is checking a box. As one of the guys who implements what happens when you check that box, I think maybe we should remove that checkbox so it doesn't mislead you. It LOOKS like it makes your system secure, right? Unfortunately, it mostly just makes your system slower. I can still see your ECB penguin. :)

There are both practical and technical problems with full-disk as opposed to per-user. The biggest practical problem is easily summarized as:
Do you want your files to be accessible to your soon to be ex- wife?
Generally, no, users should not have access to another user's files. When your visiting step-brother asks to borrow your laptop, he should not be handed an unencrypted copy of all of your personal and business files.

There is also a fundamental technical problem with full-disk encryption such that full-disk can either either be weak, or ridiculously slow, in most cases. It has to do with what are called "cipher modes". ECB is reasonably fast, but provides little security. CBC is secure, but modifying one sector requires updating every sector on the disk which follows it (meaning it takes a few minutes to save 1KB). Other modes are in between the two. We think that we *might* have that problem beat with a new approach, but I don't trust it yet.

> If you need to decide what ends up in your backup, you may forget something important. If you backup everything, you will have everything and cannot forget something important. The same applies for encryption.

That's absolutely true for backup, definitely. The only backup systems I recommend backup the whole damn machine. The system I designed makes *bootable* backups, that can be booted in-place as virtual machines. For encrypting and otherwise securing confidential data, there's a fundamental conflict between availability vs confidentiality and integrity. You may want to make your mp3 files openly available on your network, so you can play them with any device in the building. You might even store them in the cloud, easily accessible over the internet. You should NOT make your most confidential data readily accessible to every device on your network, including your IP camera and other cheap IoT devices with a thousand vulnerabilities each. If you're serious about security, you DO need to think about which items should be easily accessible to everyone in the company/house and which should be locked down tight.

I'll give you an extreme example of identifying the most confidential data and a very common example of failing to do so. The Coca-Cola company has perhaps a million documents that shouldn't be published on their web site, documents for employees only. Only their 146,000 employees have access to those documents, because they have some protection. (And probably anyone who cares to look can findleaked copies online). Coca-Cola also has the secret recipe for Coke. The secret recipe can't be found on Google only because Coca-Cola properly identified their most critical assets and protects them accordingly. The secret recipe isn't protected in the same way that the campus wifi password is protected. Coca-Cola has done it right.

On the other hand, it's common for web forums and similar software to treat each user's user name and password as confidential information. The security of these forums assumes that a bad guy doesn't know the target's username or password. They then expose the username via a search function or something similar. So it's a "secret" in that you assume that bad guy doesn't know it, but then it's not so secret when you display it on their list of posts. I pointed this out to the developers of two such systems and they didn't quite "get it". One understood a couple years laterb when there was a rash of attacks exploiting this problem. The second understood when I demonstrated that I could log in as admin. It *is* important to be clear on which data is public, which is "not shared", and which is secret.


Comment Re:Important milestone (Score 0) 115

Google's AI is literally leaps-and-bounds ahead of the game in that respect as the search space is so much unbelievably huger than chess that chess is laughable in comparison.

Most people are too nice to point this out, but what you just wrote here amounts to waving a bright red "I'm an idiot" flag.

Consider this: the search space of Go 25x25 is so much unbelievably huger than Go 19x19 that Go 19x19 is laughable in comparison.

But wait, I'm not done.

Consider this: the search space of Go 37x37 is so much unbelievably huger than Go 25x25 that Go 25x25 is laughable in comparison.

Just two strides, and I'm already breaking into a Cantor.

Consider this: the search space of AES 512 is so much unbelievably huger than AES 256 that AES 256 is laughable in comparison.

Are you still laughing?

Check out Game complexity. By your chosen criteria, Connect6 19x19 two decimal orders of magnitude more manly than mere Go.

Really? That's the standard you judge by?

Comment Security 102, chapter 1 - Risk Analysis (Score 4, Insightful) 242

If you go a bit beyond the corporate-mandated annual security training, most information security curriculum says that step one is identifying the assets at risk and their value. It would be silly to spend $50,000 turning your garage into a vault to protect a $15,000 car, and similarly for information security the value of the asset determines the maximum effort you should put into protecting it. This not only avoids wasting more time/money/hassle than the asset is worth, but it allows you to spend your efforts on the most valuable assets. Any time/money spent on a low-value asset is time NOT spent protecting a higher-value asset.

The identity of your favorite gaming site is worth about 5 cents US, so it is error to spend more than 5 cents worth of time trying to protect that information.

Additionally, in most cases it is better to protect and encrypt data on a per-account basis, for both technical and practical reasons. On a laptop, that means you encrypt the home directory, not the system. Multiple user logins have separate encryption, and one account can't access the encrypted files of another account. If you want to take it a step further, you can have a work account on the machine and a separate account for checking personal email, etc. Along with the obvious security benefits, that avoids having the browser or search engine auto-complete a URL based on *personal* browsing history in the middle of a presentation.

Given per-account security, a guest account with restrictions on it is quite feasible, and a theif would likely click the guest account.

Comment Per-account encryption is often better than full-d (Score 1) 242

In many cases, it is better to encrypt files for each account separately, rather than full-disk encryption. This is partly because most full-disk encryption sucks in one of two ways. (Google "ecb penguin" for an example.)

Along with avoiding technical problems with full-disk encryption modes, this improves security because the user of one account can't access files owned (and encrypted) by another account. You can even have a "guest" account for a houseguest to use, and guest can't access your files.

Since you have a guest account anyway, the guest account might also be configured appropriately given the knowledge that a thief might one day use it.

Slashdot Top Deals

The number of arguments is unimportant unless some of them are correct. -- Ralph Hartley