Unless there is market power and exploitation by monopolists.

Which just so happens to be the overarching principal of Matt Stoller's life's work.

In the long run, sure.

But as the saying goes: in the long run we’re all dead.

Everyone knows that a good reimplementation of CUDA will make them trillions but despite 15 years, no one has made anything even close to it.

There’s profit to be made and no entrants. How can we explain this? If your answer is “AMD et al tried” you aren’t paying attention to just how pathetic and shallow their efforts have been and still are.

Every GPU vendor knows that actually supporting a reimplementation of CUDA for their hardware immediately signs them up for two problems:

1. Protracted lawsuits with the (now) second-most-valuable company in the world. 2. Any significant customers being loudly reminded by Nvidia that the EULA for CUDA tools prohibits their use on non-Nvidia hardware.

The problem is not the technical challenge of reimplementing the CUDA API or tools. The problem is the users want CUDA, not something that looks and behaves like CUDA but requires them to load a different set of libraries.

These days, you don't really need to support CUDA, just PyTorch gets you most of the way there.

You have Google TPUs, Amazon Inferon, a ton of smaller players that aren't quite getting traction - but there's a ton of investment there.

In addition to the other response (risks), the economy is not a stable homogeneous field. None of the economy is a stable homogeneous field. Instead it's a bunch of discrete individuals living their life and each focusing on just a few things.

And this goes both ways. For a competitor to form requires one small team of founders who are insane enough (in a good way) to believe they should try it, and a funding source that's ready to take the gamble on/with them. So you might get a fantastically successful company "out of nowhere" (Amazon, say). Or you might get a field where nobody is at the moment on hand to try it. Or several do try it and fail silently and you never hear of them. They are individuals, not some mathematical process.

See also, even "efficient market" does not refer to THAT.

Eventually maybe but the Capitol investment is high and only a few firms are in the market place who are capable of playing. If they enter they expand they will decrease their own margins and that’s generally not desirable. Why take risks if you don’t have to and you can live out your tenure as ceo collect big bucks and Retire without any. Everyone wants to run their business slowly into the ground US steel and K-mart style

This here. We see this cycle quite often. Some economic crisis in a product hits and the price shoots up. Other entrants raise capital for the rather expensive factories required to make the product in its modern form. By the time the factories come on line, the entrenched interests fat with profits plunge prices very low and for some time driving out the new players and then you see prices rise back to higher levels and maintain that price.

It seems to many people don't realize it's not about making and selling the product. It's about the economic/money game. It's not about selling a car, it's about being a bank giving financing. It's not about selling a retail product, it's about credit cards. It's not about refining mined products, it's about controlling the willingness to invest in the sector.

Not saying it's wrong but that seems a strange. Around here there are still small mills - some with strange strategies where they will drive lumber large distances. Although they do specialize in this or that product. And seem profitable even though they are small. Meaning that the market can perhaps be entered with a small mill.

But in many economic fields there are unstable states where all producers can be worried but happy not to expand and just raise prices - as long as nobody else expands. "Eventually" can be a long time.

It's not about the size of the mill, but the capital cost to build a new one. No one is willing to invest the capital in building new mills because the old ones can always undercut their prices and kill them long before they turn a profit. It's the economic equivalent of a prisoner's dilemma.

That works against a large new entrant. No industry is going to lower their massive collective profit just to shoo away a small new entrant.

Lumbermills are a local industry since transportation is a very significant fraction of the final cost. No one is shipping wood from Oregon to a mill on the East coast to sell to a construction company in California, so every mill is very much sensitive to local competitors.

Ironically it's the larger mills that are insensitive to this dynamic, since they supply huge national customers like the big box stores or export to international markets. If Home Depot or Lowes decide their suppliers need a new mill, they finance it and it gets built.

This is what suprised me when I did a (very cursory) look in there. It seems some mills do ship back and forth quite a ways along the West Coast.

And I don't know that Home Depot and Lowes have that much of the market. They are not the ones that serve large construction projects. But if they did, it would be enough of a long term market that they could influence prices by funding new mills in exchange for more say into the price, a share of any profit, both, whatever. If they had this capacity, they could solve this funding issue.

Natural monopolies due to high capital costs are businesses like TSMC and ASML - you're not going to get a "small new entrant" with 4nm lithography for making high-end microchips.

TSMC, ASML are as extreme as it gets. You might as well ignore them completely in this thread.

There are local versions - a veterinary emergency hospital with veterinary oncologists and cardiologists in a smaller city without other metros for hundreds of miles, for example.

That’s an example where the market can only support one of there was ever a competitor it would be a question of who runs out of cash first.

Which is why no smart competitor would enter into that market - you'd never make back the profits from the price war.

Investment capital for large projects is why the stock market was invented.