add arrow-down arrow-left arrow-right arrow-up authorcheckmark clipboard combo comment delete discord dots drag-handle dropdown-arrow errorfacebook history inbox instagram issuelink lock markup-bbcode markup-html markup-pcpp markup-cyclingbuilder markup-plain-text markup-reddit menu pin radio-button save search settings share star-empty star-full star-half switch successtag twitch twitter user warningwattage weight youtube

Two-way competition

nostalgia2302
  • 20 months ago

Hello,

I'd like to discuss the fact that we live in a world where in many areas there isn't just enough competition and I don't understand why so many new startups fail.

GPUS: AMD vs NVIDIA

Mobile OS: Android vs iOS

Desktop OS: Microsoft vs Apple

Microchips: AMD vs Intel

Carbonated drinks: Coca-Cola vs Pepsi

Comics: Marvel vs DC-Comics

Shipping: UPS vs FedEx

Credit Cards: Visa vs MasterCard

I could go on, there are many business rivalries but it always seems like one of the competitors always arrive to put a stop to the other, to avoid those companies from creating a monopoly. But that's it. Beyond price control for the consumer, innovation only goes two ways which makes technology progress pretty slowly. There is nobody else doing things that those two companies don't do, in order to increase diversity. Or the alternatives are usually pretty bad or worthless.

But to make a point, let's use G-Sync vs Freesync and AMD vs Intel as examples:

  1. Why has nobody created an open source, hybrid adaptive sync technology that works with both GPU's?
  2. Why has nobody created an Hybrid Microprocessor that can work both in AM4 and LGA 1151 sockets?
  3. Why has nobody created an Hybrid CPU that can switch between Multicore and Single core performance (like, an adaptive CPU)? Think of an Hybrid car which can swap between gas engines and electric engines.

FirefoxOS failed. Opera Browser failed. Ubuntu Touch failed. Blackberry OS failed and they were forces to switch to Android. Other comic companies are basically unheard off. Linux only has like 2% of the Desktop market. Once two businesses have established their domination and created dependence, it on their products, it seems like it's impossible to innovate and change things.

For example, Mobile OS seem to have a dependence on Android apps. At some point, FirefoxOS and Ubuntu Touch announced some kind of compatibility with Android apps, either with an emulator or natively. Sailfish OS needs a proprietary compatibility layer to support Android apps. Even ChromeOS did it.

Basically, it's like saying that if the OS cannot support Whatsapp, nobody is going to use it. Therefore it fails, and its development is abandoned. It's impossible for a company to create a successful, standalone product without having to rely on the competition.

Comments

  • 20 months ago
  • 6 points

Why has nobody created an open source, hybrid adaptive sync technology that works with both GPU's?

AMD did. Freesync as far as I am aware is open to be implemented by any GPU manufacturer. nVidia just wants to sell more stuff. But they are a company. And gsync is better in most aspects, but both work well.

Why has nobody created an Hybrid Microprocessor that can work both in AM4 and LGA 1151 sockets?

Because that would be almost impossible. Two different interface methods, two completely different chipsets designed for two completely different architectures. Two completely different RAM controllers designed to work with the chipsets and CPUs that can't simply be switched. And a lot of other reasons. It just wouldn't be practical or even reasonable cost wise.

Why has nobody created an Hybrid CPU that can switch between Multicore and Single core performance (like, an adaptive CPU)? Think of an Hybrid car which can swap between gas engines and electric engines.

This doesn't really make sense. All modern desktop CPUs have the ability to either maximize performance on one core, or all cores. Explain this a bit more and it might make sense.

  • 20 months ago
  • 1 point

I know. The way I asked it was pretty weird. But I'm ready to get into detail so you can understand.

In general terms, Intel Single Core is more powerful than AMD's and AMD's multicore is more powerful than Intel's.

Now imagine you have a basic microchip. But such a chip is not engineered to be "multipurpose" only like Intel's and AMD's. It would be a chip that could run between 1ghz and 5ghz, and in the BIOS config you could choose between "Single Core mode", "Multi Core Mode" and "Standard Mode". You could choose to activate and disable the number of cores and their clock speed from a software. "Gaming mode" would disable all cores except 4, and run them at 5GHZ. 3D Rendering mode would activate all cores and give them the most powerful single core IPC as possible. Video rendering mode would activate all cores at a moderate clock speed.

It would have to be a chip, a die with two different architectures inside it where you can adapt the chip to what you want to do.

It's like creating a chip with two different profiles/architectures inside. One with Intel-like Single core performance and other with AMD's multi-core performance, and you would be able to choose which architecture you want to enable when booting up.

I probably just made this more complex but anyways, let me know if you get the idea

  • 20 months ago
  • 4 points

But thats not how it works. Core for core Intel has better multicore performance. AMD has price to performance tho.

Also CPUs automatically adjust thenselves to prioritize single or multicore performance.

And the ability to disable cores to be able to increase the clockspeed and decrease heat on other already exists in the BIOS of most modern motherboards. I think what you are missing is that all the things you just described modern chips do.

Also having a chip with two different architectures just wouldn't make sense.

  • 20 months ago
  • 1 point

I guess you're right. I was trying to explain a weird idea. I probably got caught in the trying to explain a very futuristic version of a processor...for things that already exists.

  • 20 months ago
  • 1 point

I think I get your underlying point. If AMD and Intel teamed up, then with some effort they could combine their technologies and make a better chip than either could alone—in the short run.

(Aside: In fact they've partly done this with Intel's x86 and AMD's x86-64. The processor industry is kind of weird in that the major competitors have cross-licensed their core technologies.)

But the combined firm would be a monopoly that might have less incentive to innovate; competition and market forces also have value to society by encouraging firms to maximize efficiency. Theoretically, in the extreme we could maximize efficiency by having one company buy up the whole economy and assign resources by command to their best uses. In practice, that doesn't work.

  • 20 months ago
  • 3 points

I don't understand why so many new startups fail.

As far as the tech examples (and some others) go, the head-start of existing companies plays a huge role. Intel/AMD/Nvidia/etc. already have the necessary infrastructure and relationships with those along the material production line for their respective product lineups. These are really quite specialized industries.

Think of all the investments and connections you'd need to make before producing a functional component (in the eyes of the end-user) outside of the big-name ones. Building something comparable to a Pentium 4, even, from raw material to final product is essentially an impossible task for one person that's new to the industry, so there'd be so many people to bring on board, so many issues to work with suppliers that may or may not be welcoming to a new start-up, and all this while the existing companies largely pursue business as usual. It would take such a huge amount of time and money for something that simply might not succeed against the industry giants.

If, rather than attempting to compete in an existing sector with well-established gigantic corporations, you attempt a start-up with new ideas or a sufficiently original expression of existing ideas, perhaps you'd have a better shot.

  • 20 months ago
  • 1 point

Nobody said that innovation comes with a price. There are many companies out there (well, not companies, individuals) with amazing ideas that seem to get lost in the sea of existing corporate domination.

Nvidia was founded in 1993 with $40,000 USD. that's about $70,000 USD today adjusted to inflation. 25 years later, they're a huge succesful company.

Nothing prevents a, let's say, 3 recently graduated computer engineers with ''insert original and attractive idea here'' to get a loan for $70,000 USD and start their new GPU company to compete against AMD and NVIDIA.

  • 20 months ago
  • 3 points

Nvidia was founded in 1993 with $40,000 USD. that's about $70,000 USD today adjusted to inflation. 25 years later, they're a huge succesful company.

If you're motivating that conclusion by reasoning that a new company could follow in Nvidia's footsteps, you'd also conclude that investing an equivalent amount into a similar company 25 years after Nvidia would result in them being 25 years behind.

Either way, I disagree with the entire premise of having a new startup develop, from today, in a manner fully comparable to the path Nvidia or ATI once took.

Nothing prevents a, let's say, 3 recently graduated computer engineers with ''insert original and attractive idea here'' to get a loan for $70,000 USD and start their new GPU company to compete against AMD and NVIDIA.

From a certain philosophical perspective, perhaps, nothing prevents that. However, 3 graduates and $70,000 isn't remotely near enough to have a hope of catching up to AMD and NVIDIA in a reasonable amount of time. The learning curve is far too steep.

When does the loan get paid back, and when do the founders start seeing proceeds? Generally, not until they either get more investments, or have a marketable product. The latter doesn't come until after they can remotely compete with AMD/Nvidia in at least some niche, if not the consumer GPU market, so for the initial phase, they'd be stuck with the former. Perhaps, if their idea is sufficiently original and attractive, big/many investors may come. Otherwise, competing against the giants looks like a losing proposition.

  • 20 months ago
  • 2 points

Not to mention that research today costs a whole lot more than it did 25 years ago...

What would it take for a third CPU/GPU company to emerge, in your opinion? a business alliance, perhaps? I remember AMD signed a deal with Intel so they could use each other's technology or something like that.

  • 20 months ago
  • 2 points

Actually, Intel keeps saying their going to enter the GPU market by 2020. Then it’ll be a three-way competition.

Intel also passes the checklist that you and Root_user have created.

  1. experience with silicon (CPUs and iGPUs)
  2. fat wallet
  3. connections in the industry
  • 20 months ago
  • 2 points

Credit Cards: Visa vs MasterCard

American Express, Capital One

  • 20 months ago
  • 1 point

Discover

Capital one is a bank that makes Visa and Master cards lol

  • 20 months ago
  • 1 point

Coopetition not Competition.

They cooperate in order to keep others out of their market, well maintaining a façade of being in competition.

  • 20 months ago
  • 1 point

If that's so, shouldn't that be illegal? that is preventing a healthy diversity in the market.

  • 20 months ago
  • 1 point

Healthy from a consumer standpoint can mean unhealthy from a manufacturer standpoint.

By spreading out to more options all you do is reduce the funding each has access to causing less progress and slower advances in each field.

Let's say you have ten companies producing CPU, they will all likely need separate sockets to protect their own requirements for hardware support, they will all need software optimization specific to their own requirements, and they all have to compete for the same money leaving each with less to put back into research and development on newer products.

Too much diversity and competition can be as bad or worse then not enough.

Flip side of that however is with only two companies they have no choice but to work together to keep the other viable.

So you see when AMD for example falls behind in one of their markets the competitor slows their advancement to a snail's pace so they can catch up.

The wild card in everything is the greedy retailer who are not held to the same standards and will gouge prices at any and every opportunity.

[comment deleted]
  • 20 months ago
  • 1 point

Once two businesses have established their domination and created dependence, it on their products, it seems like it's impossible to innovate and change things.

In economics, this is related to the concept of natural monopoly (https://en.wikipedia.org/wiki/Natural_monopoly). It does seem that there are a lot of natural monopolies in tech, due to the need for compatibility. Governments can try to mitigate natural monopolies by imposing regulation or providing subsidies; of course, this intervention can have unintended consequences.

innovation only goes two ways which makes technology progress pretty slowly.

This is not necessarily true. In the absence of a natural monopoly, the incumbent duopolists may still be forced to innovate, not by an actual competitor, but rather by the threat that if they stop innovating, a new competitor will emerge and leap-frog them. In fact, this can even occur if there's an actual monopoly (only 1 company in the market). https://en.wikipedia.org/wiki/Contestable_market

There's an argument that a duopoly can provide better competition. Suppose there are liquidity constraints (for various reasons, it may be expensive for companies to borrow money, even if their proposed use for the money is theoretically productive). Then in a duopoly with 2 large companies, if one company stops innovating, the other big company can use its spare resources to advance. In contrast, if there are 10 small companies, none of them may have enough free resources at its disposal to easily do this.

[comment deleted by staff]

Sort

add arrow-down arrow-left arrow-right arrow-up authorcheckmark clipboard combo comment delete discord dots drag-handle dropdown-arrow errorfacebook history inbox instagram issuelink lock markup-bbcode markup-html markup-pcpp markup-cyclingbuilder markup-plain-text markup-reddit menu pin radio-button save search settings share star-empty star-full star-half switch successtag twitch twitter user warningwattage weight youtube