3 ms·

As much as I really loathe Qualcomm, they placed an enormous existential bet on QAM and that they could make the linearity requirements work. For a long time,

by bsder 4y ago

As much as I really loathe Qualcomm, they placed an enormous existential bet on QAM and that they could make the linearity requirements work. For a long time, the power amps to make that work had an absolutely abysmal yield and only a single company (Rockwell Defense) could fab them. The company walked a tightrope and could very well have died. This is the kind of tech advancement that absolutely deserves some level of patent protection. The fact that the patent system allows this to be parlayed into a multi-decade shakedown is different from whether Qualcomm should have been granted patents initially.

sydthrowaway 4y ago
What's the next future in modulation schemes?
- femto 4y ago
  Massive MIMO. Arguably a modulation scheme if the array of antennas is treated as a whole.
xoa 4y ago
I think the basic problem was in their abuse of the standards process. Having patents on your own inventions for a limited time, for hardware at least [0], has strong arguments. But it doesn't typically exempt one from general antitrust law as far as colluding with others to corner a market. On the other hand, there is an unavoidable natural physical limitation when it comes to the viable electromagnetic spectrum, and huge win/wins all around in everyone being standardized across carriers and devices for how they communicate. The spirit of the SEP process was supposed to be essentially a quid pro quo: everyone can volunteer to join and then have their IP used in a way that would normally be illegal (all players leveraging various government granted or built monopolies to essentially force the entire world to use and pay for certain IP), and thus get a massively larger market for their products, but in exchange have to give up the normal full levels of control a patent would afford by offering RAND terms to all on the value of those specific contributions. It's a bet that more money can be made from a guaranteed smaller cut on billions of devices vs a large cut on ones own vertical stuff (and possibly being frozen out of usage in certain markets entirely). Qualcomm though leveraged that to extract a % of everything else in the device too using this multi-level monopoly. That wasn't OK. If from the very beginning they'd wanted to just demand everyone else not use any of their patents except by specific licensing with them, no inclusion in any standards, that'd be honest enough. Or if they'd gone the other way and full RAND that too. But they wanted to have their cake and eat it and that should have been stomped on by regulators right off frankly. ---- 0: I don't think software patents should exist at all because the core investment of software is more than adequately protected by copyright.
- bsder 4y ago
  > Qualcomm though leveraged that to extract a % of everything else in the device too using this multi-level monopoly. That wasn't OK. Except that everybody could have used some other technology in the standards, but they didn't want to make the bet and drive it through because the bet was risky. And, if it failed, somebody in a big semiconductor company in a nice comfortable position making a lot of money was going to get fired. I can also turn your statement around into "Qualcomm engineered a working system while their competitors all designed bits and pieces and never put the whole thing together"--which is true, by the way. I watched some of these idiotic competitors to Qualcomm build single pieces and then not cooperate with one another because they were afraid the other guy might make a little more money than they did. Qualcomm was the beneficiary of 1) a too-long patent grant and 2) a bunch of cowardly dipshits running the competing semiconductor companies. I agree that patent grants are far too long--this is the original sin that continues to haunt us even today. If stuff like software patent grants were only like 3-5 years, there would be far less opposition and there would be far fewer problems. However, I also have negative levels of sympathy for the "competitors" to Qualcomm who were whining incessantly but didn't have the brass to actually take them on. Qualcomm was very vulnerable for a very long time, but their "competitors" were a bunch of Keystone Kops who couldn't help but shoot themselves in both feet and then the head.
stefan_ 4y ago
Sure, and you should get some money for your tech in that cellphone chip. But that doesn't mean you should be able to extort some % of a car just because it's got the chip in it.
- daniel-cussen 4y ago
  I'm not sure. It's kind of the same thing. If they can't make the car any other way because they're not capable of inventing equally good tech, meh. Like Apple allegedly did PR stunts to slander Qualcomm, to make them look bad despite creating incredibly good tech for the baseband, without which you couldn't eg have an app store because the apps wouldn't have the same bandwidth. No iPhone without it. So they pitched the jury emotional appeals, like "does this seem fair to you?" with Qualcomm's demands. They just couldn't get the tech working, that was the real problem. Qualcomm had a legitimate inventor's monopoly from my perspective. And secondly, in terms of algorithms, Qualcomm actually hires algorithm engineers whereas Apple appears not to as much. Qualcomm, Texas Instruments, Intel, Synopsys, AMD, ARM, those guys hire algorithm designers the most. Well I suppose Apple must because they make chips, so Pennsylvania Semiconductor (IIRC). The difference between algorithms is much greater than the difference between eg mechanical engines, the efficiency improvement scales up, so a 10% difference at one scale becomes 40% if you double the scale. My point is algorithms are highly elitist and there will ordinarily be one design downstream that everyone wants, and it will be highly protected by IP law. The alternative is to come up with a better one independently, nobody's stopping anybody from doing that, but this turns out daunting and expensive.

by bsder 4y ago

sydthrowaway 4y ago

What's the next future in modulation schemes?

femto 4y ago
Massive MIMO. Arguably a modulation scheme if the array of antennas is treated as a whole.

xoa 4y ago

I think the basic problem was in their abuse of the standards process. Having patents on your own inventions for a limited time, for hardware at least [0], has strong arguments. But it doesn't typically exempt one from general antitrust law as far as colluding with others to corner a market. On the other hand, there is an unavoidable natural physical limitation when it comes to the viable electromagnetic spectrum, and huge win/wins all around in everyone being standardized across carriers and devices for how they communicate. The spirit of the SEP process was supposed to be essentially a quid pro quo: everyone can volunteer to join and then have their IP used in a way that would normally be illegal (all players leveraging various government granted or built monopolies to essentially force the entire world to use and pay for certain IP), and thus get a massively larger market for their products, but in exchange have to give up the normal full levels of control a patent would afford by offering RAND terms to all on the value of those specific contributions. It's a bet that more money can be made from a guaranteed smaller cut on billions of devices vs a large cut on ones own vertical stuff (and possibly being frozen out of usage in certain markets entirely). Qualcomm though leveraged that to extract a % of everything else in the device too using this multi-level monopoly. That wasn't OK. If from the very beginning they'd wanted to just demand everyone else not use any of their patents except by specific licensing with them, no inclusion in any standards, that'd be honest enough. Or if they'd gone the other way and full RAND that too. But they wanted to have their cake and eat it and that should have been stomped on by regulators right off frankly. ---- 0: I don't think software patents should exist at all because the core investment of software is more than adequately protected by copyright.

bsder 4y ago
> Qualcomm though leveraged that to extract a % of everything else in the device too using this multi-level monopoly. That wasn't OK. Except that everybody could have used some other technology in the standards, but they didn't want to make the bet and drive it through because the bet was risky. And, if it failed, somebody in a big semiconductor company in a nice comfortable position making a lot of money was going to get fired. I can also turn your statement around into "Qualcomm engineered a working system while their competitors all designed bits and pieces and never put the whole thing together"--which is true, by the way. I watched some of these idiotic competitors to Qualcomm build single pieces and then not cooperate with one another because they were afraid the other guy might make a little more money than they did. Qualcomm was the beneficiary of 1) a too-long patent grant and 2) a bunch of cowardly dipshits running the competing semiconductor companies. I agree that patent grants are far too long--this is the original sin that continues to haunt us even today. If stuff like software patent grants were only like 3-5 years, there would be far less opposition and there would be far fewer problems. However, I also have negative levels of sympathy for the "competitors" to Qualcomm who were whining incessantly but didn't have the brass to actually take them on. Qualcomm was very vulnerable for a very long time, but their "competitors" were a bunch of Keystone Kops who couldn't help but shoot themselves in both feet and then the head.

stefan_ 4y ago

Sure, and you should get some money for your tech in that cellphone chip. But that doesn't mean you should be able to extort some % of a car just because it's got the chip in it.

daniel-cussen 4y ago
I'm not sure. It's kind of the same thing. If they can't make the car any other way because they're not capable of inventing equally good tech, meh. Like Apple allegedly did PR stunts to slander Qualcomm, to make them look bad despite creating incredibly good tech for the baseband, without which you couldn't eg have an app store because the apps wouldn't have the same bandwidth. No iPhone without it. So they pitched the jury emotional appeals, like "does this seem fair to you?" with Qualcomm's demands. They just couldn't get the tech working, that was the real problem. Qualcomm had a legitimate inventor's monopoly from my perspective. And secondly, in terms of algorithms, Qualcomm actually hires algorithm engineers whereas Apple appears not to as much. Qualcomm, Texas Instruments, Intel, Synopsys, AMD, ARM, those guys hire algorithm designers the most. Well I suppose Apple must because they make chips, so Pennsylvania Semiconductor (IIRC). The difference between algorithms is much greater than the difference between eg mechanical engines, the efficiency improvement scales up, so a 10% difference at one scale becomes 40% if you double the scale. My point is algorithms are highly elitist and there will ordinarily be one design downstream that everyone wants, and it will be highly protected by IP law. The alternative is to come up with a better one independently, nobody's stopping anybody from doing that, but this turns out daunting and expensive.