If you can read this, either the style sheet didn't load or you have an older browser that doesn't support style sheets. Try clearing your browser cache and refreshing the page.

(Phys Org2)   Self-healing computer chips invented. Take that, Dave   (phys.org) divider line 28
    More: Cool, Caltech, computer chips, Air Force Research Laboratory, actuators, amplifiers, battery powers, electrical engineers, microseconds  
•       •       •

2509 clicks; posted to Geek » on 10 Mar 2013 at 11:43 AM (1 year ago)   |  Favorite    |   share:  Share on Twitter share via Email Share on Facebook   more»



28 Comments   (+0 »)
   
View Voting Results: Smartest and Funniest

Archived thread
 
2013-03-10 10:10:21 AM  
 FTFA: the chips automatically developed a work-around in less than a second.

To me, at least, that isn't self-healing. That's like being blinded and learning to navigate by sound. You're still blind. You didn't regrow your eye. You just learned another way to do what you used to do.

It's amazing they can get a chip to work around a fault, don't get me wrong. But true self-healing chips aren't a reality yet.
 
2013-03-10 10:29:33 AM  
My ship runs on a carbonated logic matrix so this news doesn't bother me.
 
2013-03-10 10:45:32 AM  
Now we';re going to have to listen to the damn computer sing "Daisy" over and over again.
 
2013-03-10 10:54:58 AM  
Among the many things I hated about the movie 2010 was the lame explanation they gave for why HAL freaked out and killed the crew.  I think it was actually way simpler than that: HAL made a mistake, and just couldn't handle the fact that he did.  It was, really, the first glimmer of actual self-awareness, of true sentience.  And he couldn't handle it, and flipped the fark out.  For all his sophisticated programming, it couldn't help him deal with that.

The whole "Well, he couldn't handle the fact he had to be honest with the crew and hide the true nature of the mission" explanation just didn't work for me.  It was too needlessly complicated.
 
2013-03-10 11:30:31 AM  
sp1.fotolog.com
 
2013-03-10 11:49:06 AM  

jake_lex: Among the many things I hated about the movie 2010 was the lame explanation they gave for why HAL freaked out and killed the crew.  I think it was actually way simpler than that: HAL made a mistake, and just couldn't handle the fact that he did.  It was, really, the first glimmer of actual self-awareness, of true sentience.  And he couldn't handle it, and flipped the fark out.  For all his sophisticated programming, it couldn't help him deal with that.

The whole "Well, he couldn't handle the fact he had to be honest with the crew and hide the true nature of the mission" explanation just didn't work for me.  It was too needlessly complicated.


I think you know what the problem is just as well as I do.  You did not smoke enough cannabis before watching the movie.

/Look Dude, I can see you're really upset about this. I honestly think you ought to sit down calmly, take a bong hit, and think things over.  I've still got the greatest enthusiasm and confidence in the mission.
 
2013-03-10 12:02:24 PM  

legion_of_doo:
/Look Dude, I can see you're really upset about this. I honestly think you ought to sit down calmly, take a bong hit, and think things over.  I've still got the greatest enthusiasm and confidence in the mission.


i.chzbgr.com
 
2013-03-10 12:15:29 PM  
Didn't Gene Amdahl do something close years ago?
Due to then then high cost of silicon chip manufacturing, he designed chips that were larger, did self testing and used redundant areas that were automagically routed to with the goal of 0% manufacturing rejects.
 
2013-03-10 12:25:58 PM  
A.M.A.Z.O?
 
2013-03-10 12:44:42 PM  
Best-case scenario: The Third Law.

Worst-case scenario:  http://www.youtube.com/watch?v=OGxdgNJ_lZM">http://www.youtube.com/wa tch?v=OGxdgNJ_lZM
 
2013-03-10 12:46:33 PM  
Everybody's dead, Dave.
 
2013-03-10 12:59:32 PM  
This sounds less like "self-healing" and more like "using digital circuitry to auto-calibrate analog circuitry", which really isn't all that new.

They also carefully omit what happens when that the controlling asic gets damaged.
 
2013-03-10 01:12:26 PM  
I always think of "HAL" when I boot up my TV and it engages with the cable system.  The circular red 'beacon' winks.

/shudder
 
2013-03-10 01:15:25 PM  
newspaper.li
 
2013-03-10 01:16:25 PM  
Seems like I read about them doing something similar to this in the early 1990's, but I don't care; it was worth it for that headline! Nice one subby.
 
2013-03-10 01:28:14 PM  
That's nothing that a vat of molten steel can't fix.
 
2013-03-10 02:07:13 PM  

trialpha: This sounds less like "self-healing" and more like "using digital circuitry to auto-calibrate analog circuitry", which really isn't all that new.

They also carefully omit what happens when that the controlling asic gets damaged.


Yeah, selecting redundant parts when the primary parts die via "actuators" - useful...but not "healing".
 
2013-03-10 02:25:47 PM  
To be incorporated into a DRM scheme in 3...2...1...
 
2013-03-10 03:26:00 PM  
Sounds like hot failover to a redundant system, implemented at the IC level.

Useful for implanted medical devices maybe -- where the cost of replacing damaged electronics is emergency surgery, in the best case -- but for run-of-the-mill consumer electronics it's still going to be a better idea to dedicate every available bit of silicon to maximize efficiency in the normal case where nobody's shooting lasers at the IC while it's in operation.
 
2013-03-10 04:14:09 PM  
25.media.tumblr.com
 
2013-03-10 04:24:34 PM  

poot_rootbeer: Sounds like hot failover to a redundant system, implemented at the IC level.

Useful for implanted medical devices maybe -- where the cost of replacing damaged electronics is emergency surgery, in the best case -- but for run-of-the-mill consumer electronics it's still going to be a better idea to dedicate every available bit of silicon to maximize efficiency in the normal case where nobody's shooting lasers at the IC while it's in operation.


A lot of voice grade telecom equipment has that capacity at the equipment level. You'll have 1x1 protection where if A fails B becomes active and an alarm sent to let the technician know that A has failed.

A lot of people don't realize how much equipment fails out there because of redundancy and protection. I've seen some pieces of equipment that spent months running on the protection because the complexity and cost of fixing the primary. The end user never knows because it's designed to give seamless service.
 
2013-03-10 06:33:13 PM  

trialpha: This sounds less like "self-healing" and more like "using digital circuitry to auto-calibrate analog circuitry", which really isn't all that new.

They also carefully omit what happens when that the controlling asic gets damaged.


Yeah I asked someone who presented something about this how it was different from adaptive or background digital calibration other than a fancy name and didn't get a very solid answer.
 
2013-03-10 07:17:11 PM  
so the impact of this should be better yields and fewer device failures in the future?
 
2013-03-10 09:08:09 PM  
Could you imagine how huge this will be for the iPad and iPod?
 
2013-03-10 09:09:20 PM  

jake_lex: Among the many things I hated about the movie 2010 was the lame explanation they gave for why HAL freaked out and killed the crew.


Nonono, that was a PERFECT explanation because it shows the fallibility of human logic versus machine logic (something that programmers have been grappling with for damn near 50 years). The guy summed it up beautifully: "HAL was told to lie, by people who find it easy to lie."

Computers do not lie nor do they understand the need to. HAL was given explicit instructions to protect the mission at all costs and then was told not to reveal these instructions to the humans. Essentially he was given really BAD and contradictory instructions. Protect the mission against all threats. The humans are the mission. But what if the humans are the threat? You can see him caught in a web of logic there like Nomad from Star Trek. So he made what he thought was the correct decision. He wasn't evil, he was just given shiatty mission perameters by idiot bureacrats on Earth.

And that is totally plausible and totally realistic situation that may actually happen at some point in the future. That is, if we're stupid enough to develop AI and then allow it to control all life support functions.
 
2013-03-10 09:47:10 PM  

eyeq360: Now we';re going to have to listen to the damn computer sing "Daisy" over and over again.


 farm4.static.flickr.com

/now with dancing!
 
2013-03-10 10:19:04 PM  
Nobody else going with this?

http://en.memory-alpha.org/wiki/Borg_cube

/You shut down one system, and another one takes over without interruption
//I didn't hate Wesley
 
2013-03-11 04:10:14 PM  

Ishkur: jake_lex: Among the many things I hated about the movie 2010 was the lame explanation they gave for why HAL freaked out and killed the crew.

Nonono, that was a PERFECT explanation because it shows the fallibility of human logic versus machine logic (something that programmers have been grappling with for damn near 50 years). The guy summed it up beautifully: "HAL was told to lie, by people who find it easy to lie."

Computers do not lie nor do they understand the need to. HAL was given explicit instructions to protect the mission at all costs and then was told not to reveal these instructions to the humans. Essentially he was given really BAD and contradictory instructions. Protect the mission against all threats. The humans are the mission. But what if the humans are the threat? You can see him caught in a web of logic there like Nomad from Star Trek. So he made what he thought was the correct decision. He wasn't evil, he was just given shiatty mission perameters by idiot bureacrats on Earth.

And that is totally plausible and totally realistic situation that may actually happen at some point in the future. That is, if we're stupid enough to develop AI and then allow it to control all life support functions.


it's like how humanity dealt with fire right? it's only scary if you don't know what you're doing, and sloppily programmed AI is pretty damn scary
 
Displayed 28 of 28 comments

View Voting Results: Smartest and Funniest


This thread is archived, and closed to new comments.

Continue Farking
Submit a Link »
On Twitter





In Other Media


Report