Economy Chinese startup DeepSeek hammers US stocks with cheaper open-source AI model

paper, gunpowder, printing, and synthetic insulin say hi..
I know about paper and gunpowder, but synthetic insulin wasn't invented by China. And lets be honest, they stumbled onto a couple things while trying to mix up a bunch of random shit in search of the "elixir of life" or whatever, but when it comes to complicated scientific discoveries, they are behind.
 
Tweektown full of sh$t the CEO said a number of times was done on older H800 cards because they could not get their hands on newer H100 cards due to import restrictions and costs. This is about amount of memory they have available to build their models. In what has turned out pretty inventive they decided not to use 32 bit accuracy but 8 decimal one. Saved them time and significant amounts of processing.
The point is, it's still all NVDA.
 
The point is, it's still all NVDA.
See my other post they where running AI models on an M4 mini and the Deepseek CEO stated that there really is no reason why they could not have used other GPUs but they had already spent the money on H800.


 
I cant say I'm surprised by this news at all. Big tech bosses have played a blinder and have been unloading big bags for a while.

The demand was overstated. I guess now all has been revealed that this huge need for all these chips and energy was overcooked.

On the flip side, Well played Jensen, Well played. <mma4>
 
See my other post they where running AI models on an M4 mini and the Deepseek CEO stated that there really is no reason why they could not have used other GPUs but they had already spent the money on H800.



Makes one wonder...
Why did they already spend the money on NVDA.
Looks like that was a good choice
 
The country that dominates in AI will dominate the world. The economic ramifications should be obvious. It's a multi-trillion dollar industry. Maybe 10s or even 100s of trillions. And will have some effect on basically every existing industry. As big or even bigger than the impact of personal computers and the Internet.

But there are also military implications too. Not just in a autonomous weapons but in the scale and speed at which new advanced weapons can developed.

I get that. I mean the reason partly it is a big story is because it's causing a down turn in the stock market but I don't think that addressed the questions I had above. My questions are if the American government is literally giving millions and billions (and who knows what other things) to these small amount of corporations and they consistently get their shit pushed in consistently by other countries why is there never a re-evaluation of the small amount of people/corporations who are getting these resources?

This isn't and won't be the last time this has happened to America and America keeps giving to the same few resources with different names. America had a lead in the auto industry for a long time. Then Japan took that information and skills made things like Toyota and have been kicking American car companies in the ass for decades.
 
Makes one wonder...
Why did they already spend the money on NVDA.
Looks like that was a good choice
They explained many employees where trained on Cuda and it was just easier for them to ramp up this largely experimental AI generative model. Instead of spending a few month learning new methods.
 
See my other post they where running AI models on an M4 mini and the Deepseek CEO stated that there really is no reason why they could not have used other GPUs but they had already spent the money on H800.



Running and developing a model are very different, you aren't going to be daisychaining thousands of mac Minis as a first choice. It'll be interesting to see if Deepseek really did train on the GPUs they said, or if they were able to get time on some less legal options.
 
They explained many employees where trained on Cuda and it was just easier for them to ramp up this largely experimental AI generative model. Instead of spending a few month learning new methods.
That is great info, thank you, friend. NVDA and Cuda!
 
See my other post they where running AI models on an M4 mini and the Deepseek CEO stated that there really is no reason why they could not have used other GPUs but they had already spent the money on H800.




You can run inference on an M4. You can't train a model like DeepSeek on one though.

Running inference on apple silicon isn't a new thing. People have been doing that with Llama since it came out. There is an entire open source ecosystem built around running Llama on consumer grade hardware (llama.cpp, ollama, etc)
 
Let's just give big tech companies a trillion more in subsidies and set their tax rate to 0%.

They clearly aren't wealthy and unburdened enough to compete with the Chinese on their second and third rate GPUs.
 
I get that. I mean the reason partly it is a big story is because it's causing a down turn in the stock market but I don't think that addressed the questions I had above. My questions are if the American government is literally giving millions and billions (and who knows what other things) to these small amount of corporations and they consistently get their shit pushed in consistently by other countries why is there never a re-evaluation of the small amount of people/corporations who are getting these resources?

This isn't and won't be the last time this has happened to America and America keeps giving to the same few resources with different names. America had a lead in the auto industry for a long time. Then Japan took that information and skills made things like Toyota and have been kicking American car companies in the ass for decades.

The US government is filled with morons who know nothing about industry & innovation. Take a look at the background & education of US politicians, 99% of them are lawyers or political science majors, there's almost no one with a STEM background. These people have no idea of how technology & innovation work, what's needed to help out the tech sector, or even how to evaluate whether their plans are working. It's basically like handing a bunch of partial differential equations to some kids who can't do calculus, it's completely outside of their understanding and they have no way to make any sense of it. You basically get a case where the blind retards are leading a bunch of vulture capitalists whose main priority is mooching off of subsidies rather than innovating.

Now, take a look at the political leadership in China & Russia. Both of them have lots of STEM graduates in their government systems and a lot of their key organizations such as the ministry of energy or industry will have a STEM guy at the top or a bunch of STEM guys near the top who are running the show. These people actually have the knowledge & background to know what they're doing and how to get things moving & fix things when they ain't working. We don't have that, all we can do is throw money at the problem and pray.
 
I cant say I'm surprised by this news at all. Big tech bosses have played a blinder and have been unloading big bags for a while.

The demand was overstated. I guess now all has been revealed that this huge need for all these chips and energy was overcooked.

On the flip side, Well fxplayed Jensen, Well played. <mma4>
Maybe you are really interested in IT market history?

For example when these programmers were assumed that are just lab technicians and in order to save money stuff like COBOL etc should be developed as wunderwaffe ...we had to see progress..

Now bubble No2 ...welcome to new illusion market.

AI etc still are dumb ML shit. Also owners will have to get paying clients, real persons willing to pay for con artists.

Had been told that with computers era number of bireucrats might be reduced and number of employees in companies offices too.
Didn't happened.
Wait for repetition.

If there isn't natural intelect, AI modules can't help.
 
The Chinese are way faster in implementation. As soon as they create or steal tech then it hits the market rather fast. They also conduct research that might be ethically problematic especially when it comes to gene therapy. One biophysicist was conducting gene editing experiments to make embryos that will be immune to infections like HIV. That trait will be heritable and pass down to their children. He was pretty much making mutants and the world made a fuss. China jailed him for a few years and he is back in the lab. Some of his current backers are American. I believe they will beat us in treatment of muscular dystrophy too.
 
I don't know where you got those reports to dispute its veracity, but there a lot of BYDs down here and I never heard of any catching fire. Sounds like FUD.
According to statistical data from China’s National Fire and Rescue Administration, the rate of spontaneous fires in NEVs increased by 32 percent in the first quarter of 2024. This means that currently, an average of eight NEVs catch fire in China every day — nearly 3,000 a year.

7 Battery Electric Cars a Day Catch Fire in China: The Most Involved Brands​

 
Maybe you are really interested in IT market history?

For example when these programmers were assumed that are just lab technicians and in order to save money stuff like COBOL etc should be developed as wunderwaffe ...we had to see progress..

Now bubble No2 ...welcome to new illusion market.

AI etc still are dumb ML shit. Also owners will have to get paying clients, real persons willing to pay for con artists.

Had been told that with computers era number of bireucrats might be reduced and number of employees in companies offices too.
Didn't happened.
Wait for repetition.

If there isn't natural intelect, AI modules can't help.

A lot of this equipment being bought by Nvidia is not being used for gen AI. It's being used for other purposes such as ad revenue and keeping users engaged which has been stated by many Big Tech firms.

So now the rabbit is out the hat and we all know the need was inflated and overstated. What next is anybody's guess.
 
Aren't those the cars that are known for spontaneously blowing up?
- No. China doesnt does Teslas.

Not against you. But do you remember why Toyota kicked american car makers asses?
Instead of innovate, american car makers did the same thing decades ago. Attacked the asian car makers.

They got beayt so badly, that even the Tumbler Batmobile engine is from Toyota!
 

Tech stocks tank as a Chinese competitor threatens to upend the AI frenzy; Nvidia sinks nearly 17%​


11-6-e1503284958109.jpg

BY STAN CHOE
Updated 6:25 PM BRT, January 27, 2025
Share
NEW YORK (AP) — Wall Street’s superstars tumbled Monday as a competitor from China threatens to upend the artificial-intelligence frenzy they’ve been feasting on.

The S&P 500 dropped 1.5%, dragged down in large part by a 16.9% fall for Nvidia. Other Big Tech stocks also took heavy losses, and they pulled the Nasdaq composite down 3.1% for its worst loss in more than a month.

The damage was focused on AI-related stocks, while the rest of the market held up much better. The Dow Jones Industrial Average rose 289 points, or 0.7%, and the majority of U.S. stocks climbed. But anyone holding an S&P 500 index fund, which are found in many 401(k) accounts, felt the pain because of how influential those tech giants have become on indexes.

The shock to financial markets came from China, where a company called DeepSeek unveiled a large language model that can compete with U.S. giants but at potentially a fraction of the cost. DeepSeek had already hit the top of the chart for free apps on Apple’s App Store by Monday morning, and analysts said such a feat would be particularly impressive given how the U.S. government has restricted Chinese access to top AI chips.
images

Skepticism, though, remains about how much DeepSeek’s announcement will ultimately shake the economy that’s built around the AI industry, from the chip makers making semiconductors to the utilities hoping to electrify vast data centers gobbling up computing power.

“It remains to be seen if DeepSeek found a way to work around these chip restrictions rules and what chips they ultimately used as there will be many skeptics around this issue given the information is coming from China,” according to Dan Ives, an analyst with Wedbush Securities.

DeepSeek’s disruption nevertheless rocked AI-related stocks worldwide.

In Amsterdam, Dutch chipmaking equipment company ASML slid 7%. In Tokyo, Japan’s Softbank Group Corp. lost 8.3% to pull closer to where it was before leaping on an announcement trumpeted by the White House that it was joining a partnership to invest up to $500 billion in AI infrastructure.

crazy-fist.jpg

And on Wall Street, Constellation Energy lost more than a fifth of its value, 20.8%. The company has said it would restart the shuttered Three Mile Island nuclear power plant to supply power for data centers for Microsoft.

All the worries sent investors toward bonds, which can be safer investments than any stock. The rush pushed the yield of the 10-year Treasury down to 4.52% from 4.62% late Friday.

It’s a sharp turnaround for the AI winners, which had soared in recent years on hopes that all the investment pouring in would remake the global economy and deliver gargantuan profits along the way. Such stellar performances also raised criticism that their stock prices had gone too far, too fast.

Before Monday’s drop, which was its worst since the 2020 COVID crash, Nvidia’s stock had soared from less than $20 to more than $140 in less than two years, for example.

A small group of seven such companies has become so dominant that they alone accounted for more than half the S&P 500’s total return last year, according to S&P Dow Jones Indices. They include Alphabet, Amazon, Apple, Meta Platforms, Microsoft, Nvidia and Tesla.

Their immense sizes give them huge sway over the S&P 500 and other indexes that give more weight to bigger companies. That’s why many 401(k) holders felt the pain of Nvidia’s drop, even if they didn’t know they owned any Nvidia, so long as they owned a fund that tracks the S&P 500.

All told, the S&P 500 fell 88.96 points to 6,012.28. The Nasdaq composite dropped 612.47 to 19,341.83, and the Dow Jones Industrial Average rose 289.33 to 44,713.58.

Brian Jacobsen, chief economist at Annex Wealth Management, suggested not overreacting to Monday’s sharp swings.


“It is possible that the news out of China could be overstated and then we could see a reversal of the recent market moves,” Jacobsen said. “It is also possible that the news is true, but then that would present new investment opportunities.”

More big swings may be ahead. Apple, Meta Platforms, Microsoft and Tesla are all on the schedule this upcoming week to report how much profit they made at the end of 2024.

The pressure is on companies to keep delivering strong profits, particularly after a recent jump in Treasury yields. When bonds are paying more in interest, they put downward pressure on stock prices. Yields have been on the rise amid a solid U.S. economy and worries about possibly higher inflation coming from tariffs and other policies favored by President Donald Trump.

So far, big U.S. companies have been reporting better results than analysts expected. AT&T became the latest on Monday, and its stock rose 6.3%.

images

In stock markets abroad, movements for broad indexes across Europe and Asia weren’t as forceful as for the big U.S. tech stocks. Stocks edged 0.1% lower in Shanghai after a survey of manufacturers showed export orders in China dropping to a five-month low.

https://apnews.com/article/stocks-markets-tariffs-trump-rates-52c54e361616509280bd2775674b6b4b
 
Back
Top