Miscellaneous News

horse

Brigadier
Registered Member
It is not sustainable.

We do not know that.

Clearly we do not.

What is not sustainable, is the money put into these projects. Capital is scarce. That is limited.

A.I. could be unlimited.

Look at what Gates allegedly said in the past, that 640K is enough memory. This probably not true, because instinctively we know that when the system expands, going to need more.

The computer from being a mainframe, goes to the desktop, to our hands. Probably going further too.

The internet, went from webpages and email, to platforms, that is still expanding.

A.I. is like the computer revolution and the internet revolution, that is the expectation.

That seems unlimited. We just do not know what is next.

Remember Jensen Huang? He wanted a graphics card, so we can play computer games! Would we have said that was unsustainable because how many people really want to play computer games? Maybe, maybe not. The truth is, we just did not know.

(I know I do not play computer games, I kicked my Pac-Man habit a long time ago.)

Therefore, that is the proposition in front of us. If A.I. is unlimited, then the demand for data centers will increase over time.

We must remember how these LLM works, that a token is generated from a matrix calculation, that will generate another token, from another matrix calculation. In other words, if we ask DeepSeek a few questions, like 3 questions, the compute required grows exponentially with every question, that is my understanding.

So if A.I. is a grow industry, which it surely seems it is, as the Chinese putting A.I. into everything, that means the compute must grow exponentially to met demand.

This is a very solid business case for expanding data centers.

And look who wants to do that, it is the insiders of the industry, in China and America. Generally it is the outsiders who are not buying the market.

That is the opposite of the 1990's tech boom bust. The outsiders were buying anything and everything internet related.

Kind of interesting.

:D
 

Africablack

Junior Member
Registered Member
If NATO didn't even call Article 5 when Russian drones were literally flying over their headquarters, you think they have the guts to call Article 5 on China because of some minerals?

Opinion articles on NYT is as far as they'll go.

Also, the article is written by Koreans, who are not part of NATO in the first place. The vassals always growl the loudest. Maybe it's time for Dear General to test missiles again.
Everyone is laying into them in the comments section, I'm almost embarrassed for WaPo. They've got to figure out that these sorts of propaganda pieces don't work like they used to.
 
Last edited:

manqiangrexue

Brigadier
Chinese modesty. Full rare earths ban can hobble the US while America has actually tried everything it can think of to hobble China and failed to even do measurable damage to a Chinese economy continuing to grow at 5%. The last thing the US can try is WWIII but China's nuclear and military buildup has basically shut the door on that as well.
 

horse

Brigadier
Registered Member
Look, financial speculation, low returns, and billion-dollar investments indicate that the billion-dollar AI bubble may burst.

Look at the level of spending by Big Tech companies; this is unsustainable in the long term.

To begin with, the enthusiasm for AI is surrounded by a technocratic, deterministic, and accelerationist ideology that guarantees that AI adoption will change everything.

This narrative has some basis in the wonders of AI – after all, anyone who has used a predictive chatbot (I, today, prefer the Chinese Deepseek for obvious reasons) knows that the technology can be an excellent tool for any type of work. But it is undoubtedly exaggerated and credulous regarding what AI can actually accomplish.

We still don't know very well what artificial intelligence will be used for in the future, and a lot of money has been spent on it.

An MIT study published in the middle of this year managed to quantify the difference between this "search" within companies that have AI projects and the real financial return. Of the 300 projects analyzed, 95% had no financial return. To reach this conclusion, the researchers also interviewed 150 directors of these companies.

Another problem is the formation of an oligarchy. This trend has raised alarm bells in recent years. OpenAI bought 10% of the chip company AMD, while Nvidia is investing US$100 billion in OpenAI. Microsoft, one of OpenAI's owners, is a client of the AI cloud computing company CoreWeave, which also has Nvidia as a shareholder. Microsoft accounts for almost 20% of Nvidia's revenue.

Of course, companies can collaborate. But what we are seeing here is the formation of an interconnected oligopoly that can stifle competition, control the future of technology, and concentrate power with little or no oversight.

Furthermore, the mere announcement of these agreements helps inflate the stock prices of all of them.

I agree with what you are saying, every single point.

But, we should take it one step further.

It is what that China watcher Louis Vincent Gave said about the US stock market.

When they say NDVA is a $4 trillion dollar company, and when they say Nvidia is a $5 trillion dollar company, and everyone knows all the money is concentrated in these tech stocks, that means one thing.

Actually it means a couple of things, that one, these big tech companies are the market now, and two, since they are the market that means they are too big to fail.

So that is what Mr. Gave believes. If the bubble burst, the US government probably should and will step in.

If they bailed out the banks in 2008, then how difficult it is to bail out strategic tech companies? In fact, they already put in a price floor for strategic minerals, and the US government is a shareholder in some big name important companies, forget who now.

The game is rigged. Stay long.
 
Top