Please create an account to participate in the Slashdot moderation system

 



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

Comment Re:No It Doesn't (Score 1) 70

This find raises no such question

I respectfully disagree. I'd argue the find suggests that it is theoretically possible to find or interpolate an intact DNA strand from a long extinct creature. That SHOULD raise the question of weather it's right to do so.

It takes time to come to a cultural consensus on these things. The right time to have the conversation is now, not when some grad student shows up at her university's ethics panel saying "I can haz baby Tyrannosaurus Rex?"

That moment is far too late and we'll have lawmakers falling all over themselves in reactionary hysterics. I've seen that movie (pun intended). Let's instead have the sober adult discussions NOW so that when the foreseeable happens we can say "Yep, we knew this was coming, and have given it a lot of thought and..."

Min

Comment Re:Technology Buzz Words (Score 1) 89

I realize I'm coming back to this thread late (was travelling, sorry), but in tech writing class we learned that the best practice was to use the pluralization that would apply if we had spelled out the acronym. So in this case "Artificial Intelligences can drive cars" would be the natural sounding construction instead of "Artificial Intelligence can drive cars" where I am referring to multiple separate code bases operating in independent instantiations, as opposed to one central AI coordinating all cars, which could be inferred if I had used the forced-singular construction you are advocating. When in doubt pick the language that is clearer over the 'right' formulation.

Anyways, just wanted to comment.

Min

Comment Re:Technology Buzz Words (Score 3, Insightful) 89

The term AI has become synonymous with "something we'll see in 30 years".

AIs can drive cars and pick people out of pictures and even go back through my google photos album and match my 7 yr old with her baby picture (regardless if it does this by inferring that I only have one kid and therefore this is probably it or through facial recognition or some combination of the two, it's a nifty trick.), and beat a world class human at not only Chess but Go. My 13 yr old self, fresh out of reading Neuromancer would have told you we'd need an AI for that.

Now that we've reached there AI has receded into the future.

I saw the presentation last year DC about what they're attempting to accomplish. They want to write code that without human assistance can analyze a binary, (a neat trick by itself with the halting problem) for weaknesses, develop a patch for them, and at the same time attack its opponents and circumvent their fixes.

The real trick will be tho, the winner from the AI contest goes against the human's at Defcon on CTF. The money is heavily against the AI, but even if it places a good showing, it'll

The other issue is that people mistakenly assume that AI means Human-Level intellect (and usually greater-then human, since most of us couldn't beat a wold class Go player.

It's also separate from consciousness.

A computer that's as smart as a barn owl is artificially intelligent (and would probably be a scary good drone pilot).

Min

Comment Re:What? (Score 1) 118

Yep. If it's actually true that you live in a society where things have degraded to the point that you can't trust the system not to be systemically corrupt, you should probably leave and go somewhere else.

There are checks in any good system, but if you can't have a reasonable expectation that two members of that system won't collude, there's not much that can be done. You have too many defectors, and the particular system you are playing in is headed a bad place and you should probably quit until it gets better.

Min

Comment Re:What? (Score 3, Informative) 118

A valid point, but not really related specifically to encryption. Once evidence of any kind is gathered, how do you know the evidence is entirely original?

I expect you meant "how do we know the evidence is unaltered." Typically a hash of the data is collected at the point of collection and stored along with other details (filename, length, date/time stamp, collector information) with the collected forensics data. So the hash value can be recomputed and verify that whatever file you're looking at is the same as at the point of collection. Additionally, the standard 'chain of custody' checks can be done to verify that that hash never changed at any point in the history of custody after it was collected. If a key is available, the defense could do their own decryption to confirm that the plaintext presented is the same as the plaintext they produce from a file with the correct hash. Min

Comment Re:Stop with the nannystate warnings! (Score 1) 141

Considering that when there's an accident people slow down almost to a halt in the hope of seeing something "traumatizing", my guess is people, at least the ones who are not actively seeking attention, are not as easily "traumatized" as you think they are.

These people are, frankly, morons. I still get nightmares from a motorcycle accident I was first on the scene of 15 years ago. (PSA: If you ride, PLEASE wear a helmet, if not for you, for the poor smuck trying to save your life.) I can't imagine what it would do to me to actually hit someone. Min

Slashdot Top Deals

Machines that have broken down will work perfectly when the repairman arrives.

Working...