That's now how modern capitalism works. You can't open a corner store, because someone is already on the corner running a store. You can't open one on the empty lot down the street, because that guy running the corner store took his profits and invested in zoning laws to prevent it. If you could, there would be a store there already. Finding an open opportunity is like finding a gold nugget. It just doesn't happen very often, and definitely not in places where people have combed for them for decades.
The same applies with NVidia. You can't start a GPU company because they are the GPU Mafia. They'll run you out of business before you even get a foothold with patent suits, licensing fees, etc, etc, etc.
That's why you have startup taxi companies literally breaking laws everywhere they go. Because the laws were, in part or entirely, set up to protect established interests.
Softbank then, is working inside this system, from the standpoint of the Capitalist... the man with all the money. They're buying up corner stores, because in the long run, it is guaranteed to pay them back. Even if poorly run, it will pay off for years because of the lack of competition. Capitalism is set up to favor the owners of capital.
Lots of people start GPU companies, the limitation is entirely the high quality bar set by the three desktop competitors. NVIDIA is taking their quality advantage and slowly adapting it to the embedded space.
Literally anyone with a couple million dollars can put together an ASIC design team in a matter of months, design a functioning GLES2 GPU which gets acceptable performance, phone up a semiconductor manufacturer, and pay them about half the money to open a run for them. But even for that low bar, there is already an economic advantage to licensing somebody else's chip. When you start getting to higher feature levels and higher performance targets, the amount of real work required to design GPUs is considerably higher.
That said, if you have a few billion dollars to pour into an R&D effort, you can definitely open up a GPU shop, and no GPU mafia is going to be able to stop you.
Apple is starting it's own GPU production. They immediately were sued by their former suppliers Imagination Technology over intellectual property rights. The gall of them. Nevermind that Apple has yet to produce anything, Imagination is suing them on the premise that it is impossible for Apple to build a GPU without infringing their rights, somehow.
Apple may win in the end, it's capital vs capital. A small startup trying to steal away that Apple contract from Imagination with a innovative incredible product though; They'd be properly crushed into dust for even dreaming about it.
This is absolutely not what happened and a completely flawed analogy. One: there's talks, not lawsuit. Two: Apple claims it did clean room implementation w/o providing any proof. Apple could hire a third party to assess that w/o any risk of leaking anything to Imagination but decided not to. Because they're Apple and they can. Three: they have the upper hand here in terms of monies that they can spend. So it's not like they're this small boy strongarmed by a bully.
>So it's not like they're this small boy strongarmed by a bully.
I agree with all three of your points. Let me take a different tack to see if this makes my point more clear. Apple sued Samsung over rounded corners. The gall of them. As if anyone can own the concept of rounded corners.
Apple shouldn't be allowed to prevent other companies from building phones with rounded corner designs.
Imagination shouldn't be allowed to prevent other companies from building phones with GPU designs.
Clearly, both things happen. There are lots of phones now with rounded corners. Lots of phones with GPUs. But well placed litigation can hobble your nearest competitor or destroy a small one. The result is massive losses of productivity that could have enriched society.
All this thanks to the Capitalist idea that someone can own rounded corners or GPU designs simply because they have the most money. Capitalism rewards those with capital, not those who work hard and add value to society. Those two things are not mutually exclusive, but contributing to society is not a requirement and frequently does not occur.
Theoretically, a single super capitalist or small group of them could corner the market on everything and be rewarded for no other reason beyond ownership. Together, they could stop all progress. With companies like Apple sitting on $200,000,000,000 overseas, but failing to ship a new "pro" desktop since 4 years ago, this theoretical situation looks a lot more real every day.
Apple have also hired many senior Imagination engineers and employees over the last few years.
From Imagination's perspective, rather than changing to different technology, Apple are trying to reimplement Imagination's technology, with ex-Imagination engineers, without paying them.
Whether Imagination are right about this or not, I don't think this is a clear cut case of patent bullying.
There's no way a patent on the general idea of an in-memory framebuffer would still be valid today, there was prior art for that in the early '80s, and it's been well over 20 years since the '80s (the earliest they could conceivably have applied for a patent on that).
They also hold the patents on simulating physics on a GPU and a billion other patents.
Same goes for NVIDIA.
AMD and NVIDIA violate each other's patents all day long and many more they don't sue each other since that would be a legal balckhole for both companies.
But they will we very keen on sueing any new player.
Intel licenses a patent portfolio from NVIDIA currently for legal protections, and Samsung and Qualcomm license their GPU IP from AMD.
Totally agree. I think the parent comment was a little heavy-handed in implying that large technology companies like NVidia are anti-competitive via political and or legal means. There certainly are markets that have these qualities... but chip manufacturing is not one of them.
The economies of scale and necessary R&D investment needed to be competitive are the primary things preventing new upstarts from making an impact. This is because the industry has matured.
Although taking advantage of the power of politicians and the state is used to further the interest of capitalists in various occasions (the GP mentioned a few), why exactly is "crony capitalism" not capitalism? From a Marxian perspective, it fulfills all of the requirements for a capitalist mode of production.
And where do people "usually" call this "crony capitalism"? In fact, I tend to see this term only echoed around libertarian (the property kind, not the social kind) corners of the web.
It is also important to remember that natural monopolies can very easily form, and that the tyranny of property ownership isn't much better than the tyranny of the state granting various benefits to property owners. Proudhon in What is Property goes into this and Oscar Wilde talks about it in his essay The Soul of Man Under Socialism.
Why is the crony vs 'not crony' distinction useful, and in what ways would a 'pure' capitalist society alleviate the problems caused by crony capitalism, including but not limited to the difficulty of acquiring sufficient capital to start a business or firm, dealing with anti-competitive behavior arising out of natural monopolies, the destitution and poverty caused by the extremely low cost of labour-power in certain parts of the world?
Please don't feel as though I'm attacking, I want to learn more about your ideas.
I've read about new libertarianism, and I don't find it convincing, especially the idea of property rights. I prefer the old libertarianism that was about people, not property.
not capitalism as it defined in pure terms, or not capitalism in terms of how it always practically plays out? Can you point to a capitalist system that doesn't function like that? Can you lay out how buying regulation and manipulating market structures is not a core part of the incentives that a capitalist system lays out, and can you point out any practical ways in which you can protect capitalism from this particular pathology?
No true Scotsman in action. If all examples of capitalism eventually devolve into crony capitalism I think it's time to conclude it's part of the inherent nature of capitalism.
In order to understand what real capitalism is we'd have to understand where money came from, and even that is up for debate. Was it inefficient bartering of primitive people? (As it turns out, maybe not). Was it financiers of conquistadors? Possibly.
> The same applies with NVidia. You can't start a GPU company because they are the GPU Mafia
Comically, be sure to let ARM know that that's how things work. Founded in 1990, they shouldn't exist today if what you're claiming is true, Intel should have wiped them out of existence 20 years ago.
And be sure to let nVidia know that's how it works, 3DFX (given their early lead) or ATI should have wiped them out of existence early on. Or perhaps any number of a dozen other early big corporate leaders in graphics technology.
And be sure to let AltaVista, Lycos, Yahoo, Microsoft, AOL, Excite, et al. know that's how it works, as they should have wiped out Google.
And be sure to let Walmart, Target, Sears, JC Penney's, B&N, Waldenbooks, Borders and a hundred formerly major / soon to be bankrupt retailers know that's how it works, they all should have wiped out Amazon if your premise were true.
And be sure to let Microsoft know that's how it works, as they should have wiped out Redhat in ~1999.
And be sure to let MySpace & Friendster etc. know that's how it works, as given their extreme scale vs Facebook in 2004, they should have wiped them out.
And be sure to let Atari know that's how it works, as they should have wiped out Nintendo.
And be sure to go back in time and let Steve Jobs & Co know that's how it works, as given Apple's early size they should have wiped out Dell, Gateway, and dozens of other IBM compatible clone makers before they took over the PC market. Since that's how things work and all.
And be sure to let Netflix know that's how it works, so two dozen other massive global media & tech companies - from Comcast to Disney - know they should have wiped them out.
And be sure to let Oracle know that's how it works, they should have already wiped out Salesforce.com (and Workday, and and and and).
AltaVista was a great (I'd argue the best) search engine up until Google. More importantly, it wasn't a marketing stunt, it was the product of a research project and some really great, bright engineers.
AltaVista was created by researchers at Digital Equipment Corporation's Network Systems Laboratory and Western Research Laboratory who were trying to provide services to make finding files on the public network easier.[3] Paul Flaherty came up with the original idea,[4][5] along with Louis Monier and Michael Burrows, who wrote the Web crawler and indexer, respectively. The name "AltaVista" was chosen in relation to the surroundings of their company at Palo Alto, California. AltaVista publicly launched as an Internet search engine on December 15, 1995 at altavista.digital.com.[6][7]
At launch, the service had two innovations that put it ahead of other search engines available at the time: it used a fast, multi-threaded crawler (Scooter) that could cover many more webpages than were believed to exist at the time, and it had an efficient back-end search, running on advanced hardware.
I can justify considering it "marketing" though. They didn't consider it a way of making money, so from a business point of view it was basically a way of showing how good their research was and how powerful their hardware was. (Compare to, say, Deep Blue, or AlphaGo today.)
And to be fair, Google didn't start out knowing how they would make money either.
Yeah; I worked on a Norwegian Internet Index, which had its own search engine (our own crawler and all) back in the mid-late 90s, so I used search engines a lot back then.
Now, from my memory, AltaVista and HotBot? were the key go to players outside of Yahoo. Google was a tiny fringe player at first, until, well of course we all know how that went. So AltaVista was definitely a key player at a very significant time in Internet history.
Which doesn't detract at all from the fact that there were a dozen major search players - including AltaVista - that Google bested to take over the market, courtesy of a radically superior product. None of that should have happened if the op premise were correct. The search market was more than saturated.
Turns out one of the many flaws in the op premise, is that you can invent a superior product, or provide a superior service, in an existing category and prosper accordingly. The iPod should have never existed otherwise, there were plenty of existing mp3 players.
It pertains to the traditional physical space as well as tech. Pizza companies as an example have been killing each other for a long time, new ones rise up routinely and displace weak competitors. If the op premise were correct, that'd be impossible. Pizza Hut & Domino's were founded a quarter century before Papa John's.
That's exactly the point: the survivors prove the op's premise wrong. If the op premise were correct, the survivors would not exist, as that's what the op is claiming (that Intel exists means ARM can't exist; that Microsoft exists, means Redhat can't exist; that Walmart exists, means Amazon can't exist; that Sears existed (ha) means that Walmart could not have existed; and on it goes). Back in reality, that's not how things work. Instead of crushing Apple, Microsoft chose to help them survive. That's one example and it demolishes the op premise completely (ie competitors don't always seek absolute dominion for numerous reasons, such as political concerns; and sometimes competitors are just plain stupid or lazy or blind).
>>like finding a gold nugget. It just doesn't happen very often
>survivors prove the op's premise wrong
Nobody said it was impossible, except you. Recharacterizing a statement so you can knock it down is called a strawman fallacy. You may have heard of them.
You actually point out many examples that reinforce my point. Extremely poorly run companies like Yahoo, Sears, AOL, they're all still around. Microsoft has been poorly run for nearly two decades. It is still here. MS long since lost the server market to Linux, but that didn't stop them from throwing wrenches by proxy with SCO. MS failed miserably trying to approach the smartphone market, but that doesn't stop them from collecting licensing fees on every Android phone sold. Oracle too is trying aggressively to collect some of that Android lucre.
That's how the system works. It's winner take all. Pareto distributions everywhere. Quality products, services, and businesses are wiped out and destroyed by the rent seekers. Long after the rent seekers ceased to be productive contributors to the system, their bloated zombie corpses carry on, sucking the life out of anything new that was lucky enough to survive.
> You can't start a GPU company because they are the GPU Mafia. They'll run you out of business...
You seem pretty emphatic that it's impossible. Also, before you call Nvidia a rent seeker you might wish to look at the amount they spend on R&D and what they're doing in self-driving cars and deep learning.
Of course, a single counter example disproves the claim that it's impossible to compete with an established business.
But more realistically, we can list some successful outliers in a sea of tens of thousands of failed newcomers. It's very difficult indeed, but not impossible.
How many chip companies have seriously threatened Intel in 35 years? Not enough that it makes sense to invest in any of them but Intel. Picking ARM would have been pretty darn lucky.
In an economy the size of the US (much less the globe), as an example, very difficult means a lot of positive outcomes.
nVidia is a positive outcome vs Intel. So is Apple (after all, they sell a couple hundred million units per year of compute devices with their own customized processors, which directly threatens Intel's prosperity, of course thanks again to ARM).
Or see: McDonalds vs the hundred new burger companies (Five Guys, Shake Shack et al) that have cropped up in the last 10-15 years.
Or: Anheuser Busch AmBev vs the thousands of independent brewers that have cropped up and taken an increasingly large slice of a formerly stagnant market (thanks to government regulations) with finite physical shelf space. Talk about a perfect example of how wrong the op premise is.
Tell GM & Ford not to worry about Tesla and the electric car revolution, because they can just kill off Tesla at any time (and yes, some things like that did occur, such as Tucker - thanks to government abuse in the market).
The difference is McDonalds haven't patented "a circular ground meat disc 1-2 cm thick on a similarly shaped bun with high fructose and vinegar sauces and a token minimum of vegetables, including lettuce, pickles and tomato". I think the point is you should only be able to sue against profit a company makes, and damages against current businesses are desirable because it shows someone competing.
To simplify everything all patents should last 5 years max.
> you should only be able to sue against profit a company makes
So nobody should ever be able to sue Amazon for infringement? Do you think other companies accountants would have trouble making profit vanish to avoid lawsuits?
They're the perfect example, because Intel was a thousand times larger than them during the 1990s and they're competitors. If the op premise were correct, Intel should have (must have in fact, if we're going by the deterministic concept put forth) been able to use any number of levers to crush them, easily.
The op premise is trivially easy to prove wrong: it requires infallibility / perfect judgment on the part of competitors. That one thing alone is enough to end the argument put forth - and there are numerous other issues with it.
How could they crush arm when arm would not be infringing intels manufacturing patents. Why do you think intel was unable to make dents into the gpu space? Why did they fail at the soc business ? Why are they having a hard time in the modem business? Why did they have to acquire altera to get in the fpga business? Most hardware companies operate as oligopolies in their fields where the players have a mexican standoff going on with patents. The threat of mutually assured destruction keeps the balance going. Any new entrant faces huge challenges since they dont have the patents to hold others to ransom.
"Intel unable to make dents in the GPU business" must come from a parallel universe. Meanwhile in ours Intel has gobbled up virtually the entire low end GPU market, and is proceeding to take over the middle ground. It's the No 1 or No 2 supplier of GPUs for PCs, depending on who you ask. It's virtually impossible to buy a laptop without an Intel GPU in it. It is widely suspected in my circle (game development) that the only thing stopping Intel from shipping a competitive discrete GPU based on their Gen design (the one that goes into CPUs) is that they get much better margins selling that production capacity as CPUs. Intel sucking out the air of the GPU market is a large part of what's driving NVIDIA into "AI" and "machine learning", and AMD into custom designs (consoles).
The same applies with NVidia. You can't start a GPU company because they are the GPU Mafia. They'll run you out of business before you even get a foothold with patent suits, licensing fees, etc, etc, etc.
That's why you have startup taxi companies literally breaking laws everywhere they go. Because the laws were, in part or entirely, set up to protect established interests.
Softbank then, is working inside this system, from the standpoint of the Capitalist... the man with all the money. They're buying up corner stores, because in the long run, it is guaranteed to pay them back. Even if poorly run, it will pay off for years because of the lack of competition. Capitalism is set up to favor the owners of capital.