Credit Image: NightCafe Studio: Link
Why does the American LLMS show bad returns for investment compared to China?
The bottleneck is real: the problem is the lack of data!
GPT 3.5 to GPT-4-Tobo was quantum leap.
GPT-4.5: The very massive computing resources (as evident from the enormous prices) but not a dangerous entry into the competition.
Grok 3: 200,000 GPU in the largest group of artificial intelligence in the world – but there is no quantum leap.
Additional improvements.
Why?
Prepare to find out why Big Tech is photographing itself in its leg with a GPU account and outperforming the definitions!
And why not only China is the winner – they are advancing on all fronts.
The form is good just like the data you provide!
It was never about the account – beyond a certain point.
Once the entire internet is consumed – copyright, other than publishing, intellectual property, free use,:
There was nothing more than the models to learn from.
Artificial data is not anywhere near the data created by humans.
Driving more account and size in the problem – and you will have AGI – is a false advertisement for more chips.
Yes, the GPU account is necessary.
But Trump’s algorithms are accountable. (What is meant)
The guide?
GPT 4.5 and Deepseek R1.
Astronomical prices, improving distant performance
GPT 4.5 has a (famous astronomical) price for its application programming interface from the icons of entering 75/1 million dollars and the output codes of $ 1/1 million dollars.
Deepseek, which competes with GPT 4.5 for almost every standard (exception is emotional intelligence) Free to use to chat.
The cost of the application programming interface (API) is an average of 82 times cheaper.
And an open source of preamble!
Yes, GPT 4.5 may be an introduction to GPT -5 –
But this still does not justify the astronomical sign.
This is not pricing: This is a laundry for money in your wallet, especially in the countries of the third world.
Yes, Deepseek has been uploaded through requests in the past, but less now.
At a free price?
I am ready to wait for advanced performance!
This is the leaders of LM Arena as of the date of this article:
With their prices, GPT must be 4.5 heads and shoulders above the rest.
Instead, a licensed MIT model (Open source and for free for commercial useLess than 30 points under it on LMSYS ARNA.
And why do you lead LLMS from the Google package?
Imho:
The large amount of Google data is among them all over the world (7 models in the best 20, more than any other company) is the main reason for its dominance!
Meanwhile, in China:
China has done what Openai has identified.
Create llms moving towards AGI and make it free and open to the entire world.
The Chinese government poured billions of dollars in the research and financing of Amnesty International for private companies and advanced research in universities.
China was a neck and a fissure with the United States with regard to research- In 2024.
Now that Trump has reduced the financing of American research and the imposition of a customs tariff that could destroy any chance for the United States in the artificial intelligence race –
China is preparing to take the pole in global domination of Amnesty International and LLM.
I am waiting impatiently (as with the rest of the world, others in fear).
China has been prevented from reaching expensive GPU chips, but perhaps the best thing for Chinese artificial intelligence effort!
It forced innovation, China reached some wonderful innovations of algorithm and efficiency.
This led to The opposite of what was supposed to be carried out by the ban.
China has become a global pioneer in the efficiency of artificial intelligence and economics of LLMS with Deepseek R1.
The algorithms are the key, not an account
Calculating things only as long as the high -quality data is available and to meet the high demand for current LLM products.
To create an original AGI, almost all artificial intelligence experts agree that the best algorithms are the key.
Learning reinforcement Perhaps the promising candidate in this field.
However, the latest studies show the need to be careful in AGI.
Google DeepMind I recently released a future offer for AGI, available at the link below:
They are actually AGI’s offer with independent agents by 2026-2027!
I agree with them.
Naval Ravikant is the right and wrong about AGI (in his interview).
(I realize that going against Naval Ravikant may be a problem, and the man is a strange legend.)
It is right in more AGI’s non -production account.
I agree with him there.
But a mistake in the appearance of ASI and AGI.
AGI will develop from independent artificial intelligence factors.
Humans will not build AGI.
Artificial intelligence will.
Likewise, ASI will not create.
Amnesty International, which arrived at Agi Will.
If ASI and AGI are left for humans, 100 years may be realistic.
With all the due respect to the Navy Ravikant (again, the man is a huge and huge legend!):
You will build artificial intelligence itself and develop into AGI and ASI in ways that humans will not understand, given enough basis and independence!
Now, with the latest procedures from the higher Potus:
China is in good position to be the country to reach AGI (and therefore ASI) first.
This is a certain prediction.
conclusion
The major technology companies in the United States have been sold on the idea of more artificial intelligence chips, and more intelligence through a set of interlocking and noise ads.
Now, they are paying the price.
The tariff for every country created using a ChatGPT formula also does not help.
Research and development in multiple sectors in the United States is a disaster on all fronts.
We all know who is responsible for all this, so I will avoid naming names.
But algorithms are the key to the development of AGI.
Funding is required for the police, whatever artificial intelligence is caused by the natural development of artificial intelligence agents.
Unfortunately, if things continue in this direction, China is scheduled to become the dominant superpower in the world.
The reason I say that Amnesty International will enable breakthroughs in multiple sectors of technology that we have not seen before.
Inflatives are not imagined by humans.
Artificial Intelligence Race is the Manhattan project in the twenty -first century.
In this current scenario – China is the sure winner.
Although I do not bear my writing directly, your support helps me to continue creating articles like these without a wall or paid subscription.
If you want articles like this one in your work and your main words of your online platforms, you can get them (not on Hackernoon, but yes on other platforms)!
Call me on:
For your article! (Prices are negotiable, and I offer equal prices in the countryside.)
If you want to support my writing, think about a contribution to Patreon on this link:
https://patreon.com/c/thomascharickal/membership
Instead, you can buy coffee on this link:
https://ko-fi.com/thomascherickal
Cheers!
No Amnesty International has been used in this article except for photos.
All links in the same article, so there is no list of references this time.
