Agreed. Their back catalog of movies can be fairly tough to stomach. But I think that's because they're getting horribly squeezed b the content providers -- so Netflix has had to produce original series to keep the subscriber base, and I'm guessing that next time those catalogs roll around for licensing, that Netflix will be much more in the driver's seat. Also, Netflix isn't renewing any deals where they don't get licenses to stream it in all countries it services at once, which I think is a smart move. But that necessarily shrinks their catalog since the content providers are used to Balkanizing their distribution as much as possible and dislike the universal deals. That shrinks US content quite a bit, but I'm ok with it -- it's the right call for Netflix.
Corporations that pass through 100% of the cost increases are either in the commodity business, or aren't in a competitive arena. There is always an attempt to trim some in order to offset the new costs, simply because you know that your competitors are looking at this as an opportunity to steal some of your customers if they happen to raise their price less.
Yes, I assume that some of the trade-off is that you don't get all of the benefits of full pulse compression that you would get utilizing another waveform. So a bit worse performance, but something that is masked. I'll have to look at the waveform and see how they're hiding some of those compression techniques inside the noise well enough that they're not obviously seen. My guess is that you could still tell that it's a radar if you had the processing power available to analyze it all.
That said, this is likely only really useful against someone whom is not highly integrated -- I mean our FCC listening stations and just general RF surveillance in important areas immediately geo-locates total noise sources if multiple receivers see the same noise source. Making it look like noise might confuse or cause issues with fielded systems that are meant to identify radars (such as some of our military pulse-finding sets), and categorize them. But you can still locate the source of the emitter fairly easily.
Of course, this is probably coming out of a phased array, so they might alter the phase and amplitude tapers to absolutely minimize side lobes and keep the signal directional in order to further prevent detection.
I think that they lost a lot of the hardcore tech crowd, and it's tough to get them back. When I started reading around the mid nineties it was full of absolute experts in their domain. When I was in college in the early 2000's and going through my coursework, I commented (via another account I lost) all the time, and had great discussions about the intricacies of processor pipeline strategies, pre-fetching, assembly code hacks, nitty gritty details of compiler optimizations, and so on. Now I just don't see a lot of those people here (and in my career, I'm not dealing in depth with those anymore, so I couldn't hold up the same level of conversation). You just don't have the base load of professors, grad students, bachelors students, etc here commenting like the old days.....not sure where they've gone, or whether the industry has just changed enough that it's all web-buzzwords and cloud stuff now and everything else is mature and put to rest (or at least there's that view, correct or not)?
I think that they need to find some way to lure back a good chunk of those people -- not too sure how, but it would be nice. Or I need to find where they went, and frequent that site also. I've tried, but nothing seems quite a good as the good old days:)
All that that tells me is that the father actively didn't want to know the truth and was willing to put up with any level of cognitive dissonance in order to make reality conform to his internal belief system. I guess that the world around him should bend to his will to not accept reality? I mean, honestly -- living under the same roof, financial support, etc. But not close enough to know what's happening in his daughter's life, a daughter not willing to talk to him, and someone else spilling the beans? Still sounds like the same effect of small town gossip to me (having come from a small town).
I had a talk with a person a while ago about that scenario where I took the Devil's Advocate side of the discussion. Is it really such a horrible thing that Target knows before the family? The lady obviously knows, and it's her secret to tell, so what's the big deal with Target keeping the secret?
I mean, before big data and big stores, the same clerk might have seen you buy the pregnancy test and then the next day see you buy prenatal vitamins. If it was a small town, even if it wasn't the same cashier, their might be enough gossip to connect the two and then they would know before pretty much anyone else. Before that, it might be your bank processing checks, or the credit card company, or whatever. That particular example wasn't super-secret stuff that only a big computer with big data could have figured out.
The point of it was that it's likely always been this way for a number of things, and is that scenario really any different than the many that played out before big data? I get that big data can do a lot more and be scarier, but this particular example just doesn't make me cringe about the power of big data...
It's really symbiotic. Search engine users are worth more from an advertising standpoint, but it's unlikely that Google could have survived and grown without also creating AdSense where they got to keep a larger chunk of those revenues. It's still most of their revenue, and it's initial revenue is what gave them the money to continue to refine their search engine and keep it from being gamed. If they were just a search engine, they likely would have died.
I just need to set up about a half dozen various companies in different locales with various privacy laws, and offer a binary signing service. They sign with their key, and then each of the various independent companies wholly owned in various jurisdictions also signs them, and only once that's happened is the signature correct for the device to authorize.
There's some M of N signers trust webs out there, as well as some block-chain signing type things. Basically implement that, make it international, have the companies be independent entities with no tie to each other or their customers. Then you would have to coerce/steal keys in a half dozen various jurisdictions to get a binary authorized.
Yup. Title is a little premature. They lit it up some places -- if they make it an official policy, and say partner with one or two of these companies to have the ability to light up their dark fiber nationwide, then the title might make sense. Right now, we're not there yet.
I think that the issue is that a lot of them were getting repetitive. I research a lot of products at work and regularly click on the ads in addition to the search results. The ads are usually pretty relevant and in some cases better than what the search terms are providing. But I was noticing that a single company's ads might be shown 2-4 times across all of the locations. I can see how that could cause issues.
Sure, both of them do... but with airlines you build the infrastructure at the end-points only. The bits which connect those end-points? You don't need to build anything, because it's just the atmosphere.
You mean other than the entire airspace spanning Air Traffic Control infrastructure that costs ~$7B per year to run? Lots of radar, stations, towers and other stuff to control things other than just the end-points.
This is interesting to me.....even if they didn't request the IP, in DOD contracts, if you spent money on it, it at least becomes FOUO. Meaning that you can't really license it. I had thought that it would be government wide -- that if government money is spent on it, even if some was chipped in from the other entities, that it becomes if not fully owned government IP, at least free for government use.....maybe the DOD had just learned their lessons earlier...
From what I can tell, they were needing to fix just a couple of bad letters in Highway Gothic to be more readable. So, why not start with that as a base?
Will it be a truck like the Model X is an SUV (aka, not really one -- can't beat it up offroad like you can a 4Runner, Highlander, etc)? If it's a real truck that can go offroad, then that would be great. But that's a big step versus where they are now with the X...
Then they'd physically broken into your home and rifled through your stuff, likely stole valuables and potentially enough PII to essentially become you (passport, SS card, birth certificate, etc). You have bigger issues than whether they can log into your gmail account or not....
For some reason Bennett thinks that the free market works nearly instantaneously, and everyone gets perfect information instantly and instantly makes the switch. The real world takes a bit longer than that for information and action to spread. It often takes multiple years before a market shakes out problems like this. This is Econ 101 stuff -- stop puzzling and read some case studies.
Pretty much all companies totally suck at M&A's. I don't know if he has much of a leg to stand on here given large company track records of small single digit successes in M&A.
They have the drawings, but that doesn't mean that they have all of the processes -- these are complex items operating at the edges of materials science. The Russians are still more advanced than the US in many metallurgical sciences. There are some alloys and specific metal grain configurations used in the RD-180 that simply no one else knows how to do but the Russian shops that build the RD-180 engine, which are under this embargo.
Cue the talk about when we were looking to resurrect some Saturn V engines, there were numerous compounds and other items that we didn't know how to produce either -- the proprietary process died with the one-man shop that knew the secret sauce and had the dialed in equipment.
I'm pretty pumped about playing with the dev kit. It has a heterogeneous memory architecture between the CPU and GPU. For lots of GPGPU applications, the latency of transfer between system RAM and the GPU can be a bottleneck. You're transferring huge chunks of data, and if you need to bounce the problem back and forth between the CPU and GPU, which is pretty common or if you have any real-time requirements, it can be a big deal. In many applications it can be 40%+ of time spent in just transferring your data back and forth from GPU to CPU.
For example, lots of people used the TK1 (predecessor to the TX1) for computer vision applications because it ran faster than the fastest GPU merely because you didn't have the memory transfer times. But the TK1 was slightly underpowered for these applications. The TX1 should close that gap, and really allow true GPU/CPU co-processing, versus shuttling around in memory.
Nvidia will be bringing heterogeneous computing to the desktop soon too -- they're already making it happen with IBM, and then their roadmap is to push it to x86 land.
I fail to see how this is the case. I can't provide bagels to anyone in the US Government during my meetings because various laws do not allow it -- it would be seen as currying favor, and they must pay for it all themselves (well, you can give them up to $50 a year or something, but that's just one breakfast+lunch out of many in a 5-day meeting on the East Coast). But you can effectively donate $MILLIONS to a politician? I say that if there truly is a 1st Amendment issue at stake here, then the various government procurement officers should at least get in on the game too. Let the graft commence!
Once you start telling people how they can spend their own money, freedom is just a joke.
You cannot legally buy votes.
You cannot legally pay to have someone killed.
You cannot legally buy another person.
Obviously freedom is just a joke and we should be allowed to do these things.
Or, one could realize that freedoms among people are various balancing acts, and that striking the right balance is a good one. I don't think that you should be able to effectively buy a politician's vote. It's corrosive to our government, and our government is whom we charge with enforcing our notions of freedom. Hence, more freedom is preserved if we restrict this one; similarly with slavery, contract hit jobs, etc.
I don't buy these reports at all. I just don't see it. I would imagine that Google would like to partner a bit closer with some of the chip vendors -- get some low power extensions added, more direct hardware accelerations of some of the effects that are done in Android, maybe help define some other extensions, etc. But I seriously doubt that they're looking to get into the chip design business. To do so they would have to buy a slew of chip designers, and we just haven't seen them hiring or acquiring in that arena.
Chip design is very hard and unforgiving. Google knows this, and can't be looking to jump into the business. They might want to help tailor something, but that would be about it...
So without a resistor how do you know whether the cable is specified to carry 500mA, 1A, 2A or 3A? They all look the same and have the same connectors on both ends. Trying to draw 3A over a 1A rated cable will cause issues for everyone involved...
Agree, which is why you're also support to support the USB Type-C negotiation protocol, which typically both ends do. There do exist some dumb chargers though, and lots of dumb implementations exist where the controller just gives whatever was asked for even if it can't supply it -- cheaper to build that way rather than spend time on pesky programming and engineering. But the big thing is that you also have to worry about the spec of the cable. Likes of USB 2.0 cables really do only support 500mA, or 1A or current. 3A would fry them, so you need something in the cable to identify that you can safely pass that much current across them. This resistor lets them know...
Slashdot Mirror
Agreed. Their back catalog of movies can be fairly tough to stomach. But I think that's because they're getting horribly squeezed b the content providers -- so Netflix has had to produce original series to keep the subscriber base, and I'm guessing that next time those catalogs roll around for licensing, that Netflix will be much more in the driver's seat. Also, Netflix isn't renewing any deals where they don't get licenses to stream it in all countries it services at once, which I think is a smart move. But that necessarily shrinks their catalog since the content providers are used to Balkanizing their distribution as much as possible and dislike the universal deals. That shrinks US content quite a bit, but I'm ok with it -- it's the right call for Netflix.
Corporations that pass through 100% of the cost increases are either in the commodity business, or aren't in a competitive arena. There is always an attempt to trim some in order to offset the new costs, simply because you know that your competitors are looking at this as an opportunity to steal some of your customers if they happen to raise their price less.
Yes, I assume that some of the trade-off is that you don't get all of the benefits of full pulse compression that you would get utilizing another waveform. So a bit worse performance, but something that is masked. I'll have to look at the waveform and see how they're hiding some of those compression techniques inside the noise well enough that they're not obviously seen. My guess is that you could still tell that it's a radar if you had the processing power available to analyze it all.
That said, this is likely only really useful against someone whom is not highly integrated -- I mean our FCC listening stations and just general RF surveillance in important areas immediately geo-locates total noise sources if multiple receivers see the same noise source. Making it look like noise might confuse or cause issues with fielded systems that are meant to identify radars (such as some of our military pulse-finding sets), and categorize them. But you can still locate the source of the emitter fairly easily.
Of course, this is probably coming out of a phased array, so they might alter the phase and amplitude tapers to absolutely minimize side lobes and keep the signal directional in order to further prevent detection.
I think that they lost a lot of the hardcore tech crowd, and it's tough to get them back. When I started reading around the mid nineties it was full of absolute experts in their domain. When I was in college in the early 2000's and going through my coursework, I commented (via another account I lost) all the time, and had great discussions about the intricacies of processor pipeline strategies, pre-fetching, assembly code hacks, nitty gritty details of compiler optimizations, and so on. Now I just don't see a lot of those people here (and in my career, I'm not dealing in depth with those anymore, so I couldn't hold up the same level of conversation). You just don't have the base load of professors, grad students, bachelors students, etc here commenting like the old days.....not sure where they've gone, or whether the industry has just changed enough that it's all web-buzzwords and cloud stuff now and everything else is mature and put to rest (or at least there's that view, correct or not)?
:)
I think that they need to find some way to lure back a good chunk of those people -- not too sure how, but it would be nice. Or I need to find where they went, and frequent that site also. I've tried, but nothing seems quite a good as the good old days
All that that tells me is that the father actively didn't want to know the truth and was willing to put up with any level of cognitive dissonance in order to make reality conform to his internal belief system. I guess that the world around him should bend to his will to not accept reality? I mean, honestly -- living under the same roof, financial support, etc. But not close enough to know what's happening in his daughter's life, a daughter not willing to talk to him, and someone else spilling the beans? Still sounds like the same effect of small town gossip to me (having come from a small town).
I had a talk with a person a while ago about that scenario where I took the Devil's Advocate side of the discussion. Is it really such a horrible thing that Target knows before the family? The lady obviously knows, and it's her secret to tell, so what's the big deal with Target keeping the secret?
I mean, before big data and big stores, the same clerk might have seen you buy the pregnancy test and then the next day see you buy prenatal vitamins. If it was a small town, even if it wasn't the same cashier, their might be enough gossip to connect the two and then they would know before pretty much anyone else. Before that, it might be your bank processing checks, or the credit card company, or whatever. That particular example wasn't super-secret stuff that only a big computer with big data could have figured out.
The point of it was that it's likely always been this way for a number of things, and is that scenario really any different than the many that played out before big data? I get that big data can do a lot more and be scarier, but this particular example just doesn't make me cringe about the power of big data...
It's really symbiotic. Search engine users are worth more from an advertising standpoint, but it's unlikely that Google could have survived and grown without also creating AdSense where they got to keep a larger chunk of those revenues. It's still most of their revenue, and it's initial revenue is what gave them the money to continue to refine their search engine and keep it from being gamed. If they were just a search engine, they likely would have died.
I just need to set up about a half dozen various companies in different locales with various privacy laws, and offer a binary signing service. They sign with their key, and then each of the various independent companies wholly owned in various jurisdictions also signs them, and only once that's happened is the signature correct for the device to authorize.
There's some M of N signers trust webs out there, as well as some block-chain signing type things. Basically implement that, make it international, have the companies be independent entities with no tie to each other or their customers. Then you would have to coerce/steal keys in a half dozen various jurisdictions to get a binary authorized.
Yup. Title is a little premature. They lit it up some places -- if they make it an official policy, and say partner with one or two of these companies to have the ability to light up their dark fiber nationwide, then the title might make sense. Right now, we're not there yet.
Other than the cost of billions of dollars of to amortize....
I think that the issue is that a lot of them were getting repetitive. I research a lot of products at work and regularly click on the ads in addition to the search results. The ads are usually pretty relevant and in some cases better than what the search terms are providing. But I was noticing that a single company's ads might be shown 2-4 times across all of the locations. I can see how that could cause issues.
Sure, both of them do ... but with airlines you build the infrastructure at the end-points only. The bits which connect those end-points? You don't need to build anything, because it's just the atmosphere.
You mean other than the entire airspace spanning Air Traffic Control infrastructure that costs ~$7B per year to run? Lots of radar, stations, towers and other stuff to control things other than just the end-points.
This is interesting to me.....even if they didn't request the IP, in DOD contracts, if you spent money on it, it at least becomes FOUO. Meaning that you can't really license it. I had thought that it would be government wide -- that if government money is spent on it, even if some was chipped in from the other entities, that it becomes if not fully owned government IP, at least free for government use.....maybe the DOD had just learned their lessons earlier...
From what I can tell, they were needing to fix just a couple of bad letters in Highway Gothic to be more readable. So, why not start with that as a base?
Will it be a truck like the Model X is an SUV (aka, not really one -- can't beat it up offroad like you can a 4Runner, Highlander, etc)? If it's a real truck that can go offroad, then that would be great. But that's a big step versus where they are now with the X...
Then they'd physically broken into your home and rifled through your stuff, likely stole valuables and potentially enough PII to essentially become you (passport, SS card, birth certificate, etc). You have bigger issues than whether they can log into your gmail account or not....
For some reason Bennett thinks that the free market works nearly instantaneously, and everyone gets perfect information instantly and instantly makes the switch. The real world takes a bit longer than that for information and action to spread. It often takes multiple years before a market shakes out problems like this. This is Econ 101 stuff -- stop puzzling and read some case studies.
Pretty much all companies totally suck at M&A's. I don't know if he has much of a leg to stand on here given large company track records of small single digit successes in M&A.
They have the drawings, but that doesn't mean that they have all of the processes -- these are complex items operating at the edges of materials science. The Russians are still more advanced than the US in many metallurgical sciences. There are some alloys and specific metal grain configurations used in the RD-180 that simply no one else knows how to do but the Russian shops that build the RD-180 engine, which are under this embargo.
Cue the talk about when we were looking to resurrect some Saturn V engines, there were numerous compounds and other items that we didn't know how to produce either -- the proprietary process died with the one-man shop that knew the secret sauce and had the dialed in equipment.
I'm pretty pumped about playing with the dev kit. It has a heterogeneous memory architecture between the CPU and GPU. For lots of GPGPU applications, the latency of transfer between system RAM and the GPU can be a bottleneck. You're transferring huge chunks of data, and if you need to bounce the problem back and forth between the CPU and GPU, which is pretty common or if you have any real-time requirements, it can be a big deal. In many applications it can be 40%+ of time spent in just transferring your data back and forth from GPU to CPU.
For example, lots of people used the TK1 (predecessor to the TX1) for computer vision applications because it ran faster than the fastest GPU merely because you didn't have the memory transfer times. But the TK1 was slightly underpowered for these applications. The TX1 should close that gap, and really allow true GPU/CPU co-processing, versus shuttling around in memory.
Nvidia will be bringing heterogeneous computing to the desktop soon too -- they're already making it happen with IBM, and then their roadmap is to push it to x86 land.
I fail to see how this is the case. I can't provide bagels to anyone in the US Government during my meetings because various laws do not allow it -- it would be seen as currying favor, and they must pay for it all themselves (well, you can give them up to $50 a year or something, but that's just one breakfast+lunch out of many in a 5-day meeting on the East Coast). But you can effectively donate $MILLIONS to a politician? I say that if there truly is a 1st Amendment issue at stake here, then the various government procurement officers should at least get in on the game too. Let the graft commence!
Once you start telling people how they can spend their own money, freedom is just a joke.
You cannot legally buy votes. You cannot legally pay to have someone killed. You cannot legally buy another person. Obviously freedom is just a joke and we should be allowed to do these things. Or, one could realize that freedoms among people are various balancing acts, and that striking the right balance is a good one. I don't think that you should be able to effectively buy a politician's vote. It's corrosive to our government, and our government is whom we charge with enforcing our notions of freedom. Hence, more freedom is preserved if we restrict this one; similarly with slavery, contract hit jobs, etc.
I don't buy these reports at all. I just don't see it. I would imagine that Google would like to partner a bit closer with some of the chip vendors -- get some low power extensions added, more direct hardware accelerations of some of the effects that are done in Android, maybe help define some other extensions, etc. But I seriously doubt that they're looking to get into the chip design business. To do so they would have to buy a slew of chip designers, and we just haven't seen them hiring or acquiring in that arena.
Chip design is very hard and unforgiving. Google knows this, and can't be looking to jump into the business. They might want to help tailor something, but that would be about it...
So without a resistor how do you know whether the cable is specified to carry 500mA, 1A, 2A or 3A? They all look the same and have the same connectors on both ends. Trying to draw 3A over a 1A rated cable will cause issues for everyone involved...
Agree, which is why you're also support to support the USB Type-C negotiation protocol, which typically both ends do. There do exist some dumb chargers though, and lots of dumb implementations exist where the controller just gives whatever was asked for even if it can't supply it -- cheaper to build that way rather than spend time on pesky programming and engineering. But the big thing is that you also have to worry about the spec of the cable. Likes of USB 2.0 cables really do only support 500mA, or 1A or current. 3A would fry them, so you need something in the cable to identify that you can safely pass that much current across them. This resistor lets them know...