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

 



Forgot your password?
typodupeerror
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. Also, Slashdot's Facebook page has a chat bot now. Message it for stories and more. ×

Comment Re:Low Interest In The Public (Score 1) 216

After the snowden reveal, I switched to it exclusively when communicating with a friend of mine. I use a really strong set of ECDSA keys I generated for us, and physically exchanged in person.

I laugh at the idea of the NSA wasting the CPU cycles needed to decode our harmless exchanges of adorable kitten pics.

 

Comment Re:It's a pain because recovery has to be an optio (Score 1) 216

a message that can be read by somebody other than the intended recipient, is not worthy of being called secure.

A message that can have the key derived from the data stream is a message that fails to prevent somebody other than the intended recipient from reading it.

The two are mutually exclusive.

Comment Re:Yellow Journalism (Score 1) 251

It doesn't matter anyway-- the statement is absurd. Humans will never be able to match a machine's pace, even with wires going into their heads. Organic neurons are PAINFULLY slow, with cycle times measured in milliseconds. By comparison, computers are measured in multiples of giga-ops per second.

A human will never be able to adapt to the real output pace of a well written program. Even WITH borg implants.

Comment Re:Ironically.... (Score 1) 251

No. Meat will never be able to match silicon. Neurons have a refractory period that is painfully slow compared to CURRENT silicon.

http://www.physiologyweb.com/l...

At most, a connected human brain would be a novel input source on a slow but wide bandwidth bus. There is only so much cortex you can put sensors on.

A true AI would be hobbled by adding a human brain.

Comment Re:New hard drive? (Score 1) 612

Ahh... Back when the platter in the drive did not have defined sector markings, and all drive seeking was done with step motors instead of voice coils...

If you were clever, and knew what the ACTUAL limits of your platter's magnetic density were, you could tweak the sector layout to get more total sectors than was advertised. (not by much, but back then every kilobyte mattered.)

of course, even then most people did not know what they were doing...

Comment Re:Big Floppy is scamming you (Score 1) 612

That was the Microsoft Distribution Format, or MDF.

There was another, less frequently used one that went to 1.62mb, instead of 1.72, but I forget what it was called.

Not all media was created equally though. Some disks were just total garbage if not formatted at the intended 1.44mb. I strongly suspect that they had uneven distribution of magnetic particles in the mylar, so that sector marking areas (when formatted to 1.44) were more reliable than data areas. When formatted with unusual sector and track layouts, these did not align anymore, and the disks behaved unpredictably. Others were just fine at even the strangest geometries. It was often a crapshoot to try formatting diskettes to MDF.

Thankfully the AOL install floppies were of the really nice, easy to format, and good at data retention type. :P Just put some tape over the hole, and peel off the label. You could get hundreds of them in the mail for free! Those were simpler times!

Comment Re:Radiation wrecks robots? (Score 2) 307

Semiconductors work by having a "tipping point" after which they become conductive. High energy electrons (beta particles) are way higher than the bandgap of basically any semiconductor. They will cause the migration of ions embedded in the semiconductor that enable it to be semiconductive.

Additionally, you have things like hot neutrons, and gamma rays. Hot neutrons will cause fission type interactions with the doping atoms embedded in the semiconductor, changing them into 'something else', and releasing lots of secondary particles in the process. Gamma rays are high energy photons, and contribute to migration in the semiconductor.

All in all, these all cause the semiconductor to disintegrate, and stop functioning.

That's why I suggested keeping all those sensitive parts OUTSIDE of the reactor containment vessel, and using a really fat, electrically shielded data cable and a fiber optic line attached to a dumb manipulator that goes inside instead. That way the electronics are shielded by the reactor containment walls, and however much dirt is between it and the exposed core material.

Comment Re:Money to be made... (Score 4, Interesting) 307

Random suggestion--

Use a trailing fiber optic pickup, with the actual CCD and robot controller hardware OUTSIDE the reactor.

Similar in concept to the imaging system used for laparoscopy.

In this case, the "mobile" portion of the robot is made using the "more radiation resistant" larger discrete components, with a fat data cable and fiber optic line dragging behind it, leading to the actual logic controller portion of the robot, parked outside.

That would help with costs, and service life of the robot. (Expensive controller hardware stays outside the reactor, only the driver part needs to be discarded as radioactive waste, and the imaging sensor array is not inside the reactor.)

Comment I tend to agree as well. (Score 5, Interesting) 231

Far too often, antivirus products follow the "cable television" market strategy:

"Yes, we know you already pay us for a subscription, but we can get so much more out of you by forcing you to see all kinds of shit you really don't want, including adverts for all our other services."

And, in the case of free antivirus, this too:

"We can see that you really dont want our full package, otherwise you would have bought it instead of opting for the free version-- but we feel compelled to try to upsell you each and every possible opportunity, and wont relent at all. We will even be really obnoxious with your notification area, and make your system play audio adverts, because that's how much we really want you to have a subscription (but see the prior market strategy-- we wont let up on the ads even if you do!)"

They invest tons of resources (both computational and time-wise) into making needlessly flashy UIs with big colorful buttons, and scary "CSI: Miami"-esque dialogs, when really--- the part that really matters-- how well they can trap execution events without bogging the system down-- seems to get nearly no love, and appears to get shittier and shittier.

Then you have Windows Defender. It's so plain, you instinctively ignore its presence. Excepting on older XP systems, (where there was a CPU utilization bug), it runs with a very modest system footprint. It does not constantly vomit spam into your system tray, and does not try to milk you for additional service agreements, or to switch to a paid version. It behaves itself very well.

If Avast or AVG behaved like that, instead of trying to be garishly tawdry and whorishly self-promoting like prostitutes, and reduced their system resource consumption habbits accordingly, they would win hands down.

But no, fleecing idiots is much more profitable.

Comment Re:Slight pet peeve of mine-- (Score 1) 43

A method similar to what is done for digital HDR recording would fix this problem, but would admittedly add tremendously to the vehicle weight.

EG, you have a single forward optic, but introduce several beam splitters, which then go to individual monochrome CCDs behind the various filters. (use a clear saphire prism beam splitter to get nIR and UV on the same source optic) You then assign each CCD a channel, and develop the image accordingly.

Comment Slight pet peeve of mine-- (Score 3, Informative) 43

NASA seems to really really love using monochrome imaging systems.

Look, I GET that this is in the outer solar system, and that the sun's light is very pitifully weak out there. I GET that. I understand that they want to gather as much light as is possible in the images.

However, monochrome CCDs dont care what frequency the photons are. As long as they can pass through the forward optics and focus, they will add to the luminosity of the resulting image. That means that dust could be very reflective of IR, or UV light, and it would have the same whiteness. Sure, you could subtract some of that out using special optics for IR and UV, and create some horrid false-color image that does not reflect reality at all, other than artificially showing where there is UV or IR reflectivity, but visible light absorbtion/emission spectra are also very useful for scientific enquiry into such objects.

Why does NASA not at least TRY to get true color images with extended exposure times?

It's been a pet peeve of my for years, and I cant be the only one. I KNOW they can do it, because Voyager took lots of true color images back in the 70s. CCD tech has greatly improved since then.

Comment Re:AI does what AI is programmed to do (Score 1) 169

Whoosh.

The program, is the machine learning conditional framework. The training is data that influences program execution. The program is what the machine follows, the behavior exhibited by the program is determined by the training data.

The problem, as you correctly put, is that we dont really have good feedback on what elements of the training data it is weighing on reliably.

The actual program is what defines the conditional framework at the lowest level. It does this faithfully. The emergent properties? That's another story. That is related to the data the AI has collected.

Comment Re:That is correct (Score 1) 169

Again, the danger is not "skynet deciding humans are obsolete", the danger is in "For some reason, our predator drones suddenly started mass murdering our own citizens when winter hit, and people started wearing full face balaclavas." -- Because the training they gave the predator drone in the middle east made it select for people wearing full face covering attire, because that was one of the easier metrics to weigh for.

Slashdot Top Deals

"If I do not want others to quote me, I do not speak." -- Phil Wayne

Working...