Forgot your password?
typodupeerror

Comment: Re:The Internet Needs More Random Data (Score 1) 336

by TheLink (#47423071) Attached to: UK Computing Student Jailed After Failing To Hand Over Crypto Keys
Or Ubuntu and other popular distro to do something like this:
https://bugs.launchpad.net/ubuntu/+bug/148440

Then it's normal for people to have encrypted stuff on their drives that they can't decrypt. And thus a "reasonable man" could not be expected to be able to decrypt such stuff even if he cooperated fully. They could be using full disk crypto with an encrypted container file that they can't decrypt. They can decrypt the first but not the second (or maybe they can - it becomes harder to tell :) ).

But once a popular OS has stuff like this by default, it's much easier for the defence to argue that you can't do it.

Of course in this case - the guy has been supplying wrong passwords, so unless you can show it was out of desperation and/or due to duress, he'd still be in trouble.

Comment: Re:Forget reading, GET AN IMPLANT! (Score 1) 83

by TheLink (#47422603) Attached to: A Brain Implant For Synthetic Memory
It's the wrong approach if you just want a prosthetic memory to help people remember stuff.

To have a prosthetic memory what you need is a computer that can remember stuff - video, audio, photos, text etc. Preferably wearable. Then what you need is to attach a device to appropriate parts of your brain that reads thought patterns that are distinctive depending on what you are thinking (elephants, purple etc). The device does NOT have to decipher or understand what you are thinking. All it needs to do is associate the stuff to be stored/recalled or even _commands_ with the thought pattern(s) you choose for it. I call these thought macros. See also: http://slashdot.org/comments.pl?sid=3478821&cid=42956909

So you capture a video/audio/picture then you assign it a thought, or "current state" of mind. If you even have difficulty rethinking[1] a thought pattern, you could search by context and time (what I stored some time ago while at home).

There may need to be training phases like in speech recognition, and it's likely to work better with some people than others.

[1] The approach the military is taking would still have problems if people can't even remember that they are supposed to remember something- so whichever approach you'd need the ability to set up "prompts" based on time and context (and brain patterns).

I believe our technology is very very far from the state where you can drop in a memory device with memories already preloaded in, and which people can use to "remember that they are to remember something" (and even if we did, it would be scary and I won't want to have it).

Because there's evidence that memories are stored differently on different people's brains - some people have a halle berry neuron: http://www.caltech.edu/content/single-cell-recognition-halle-berry-brain-cell
http://phys.org/news4703.html
Seems to me to be a bit like a Bingo hall where a neuron yells bingo when it recognizes what the "announcer reads out". And the thing is those neurons aren't in the same place for everyone, they might not even be present for everyone, and one neuron might yell bingo for slightly different things (in one person they might have a neuron that goes bingo for Jennifer Aniston when it sees Jennifer Aniston + Brad Pitt, in another person it might not go bingo for the couple).

Which is also why I think that it's delusional for people to believe we'd soon be able to transfer our minds to other machines. You can transfer something, but it'll be far from everything.
User Journal

Journal: Circular Refuge on reddit 4

Journal by Some Woman

It's a happening place. There are upwards of 3, maybe 4 posts a day!
You should join us, if you like.

http://www.reddit.com/r/CircularRefuge
(message mods to join; can't let the riffraff on reddit in! Just our very own special riffraff.)

Comment: Re:No they're not (Score 2) 63

by TheLink (#47394421) Attached to: Study: Whales Are Ecosystem "Engineers"

I don't see anything new or interesting in the articles to consider it a "discovery of a way" (e.g. http://www.uvm.edu/~uvmpr/?Pag... )

In contrast this is a better article with more detail on how whales could _actually_ affect ecosystems significantly: http://www.newscientist.com/ar...
And that's a 4 year old article.

news: gotcha

Working...