Sam Altmanthe executive president of OpenAi, the company to which the Chat gpt, praised the model R1 of Deepseek. This is part of an emerging company of China that showed great power for the Technological development in terms of Artificial intelligence.
Depseek is a low -cost AI model, which monopolized worldwide attention about a month ago, when he affirmed in an article that the training of his model Deepseek-V3 required less than U $ S6 million in calculation power using chips NVIDIA H800 of less capacity.
For its part, Depseek-R1, the model praised by Altman and launched last week, is Between 20 and 50 times more accessible to use that OPENAI’s O1 model. However, it also depends on the task that is performed and what is indicated to the chat.
Sam Altman’s praise
“Deepseek R1 It is an impressive modelespecially for what he is able to offer for his price, “Altman said in X last Monday.
“But, above all, we are excited to continue executing our research map and we believe that More computing is now more important than ever To succeed in our mission, “he added.
However, he never left the competition: “Obviously We will offer much better models And besides It is really stimulating to have a new competitor. We will launch some new ones. “
“The world will want to use a lot and really will be surprised by next -generation models They will come, “said the American tycoon.
Deepseek and its impact on the world market
The appearance of Deepseek brought doubts about the reasoning that underlies the decision of some US technology companies of allocate billions of dollars to investments in artificial intelligenceand the actions of several great technological ones were affected.
In that aspect, Nvidia registered one record loss of US $ 593,000 million In its market value in a single day, the largest of all companies in Wall Street. This is also linked to the fact that Chinese technology needs less powerful chips to work.
Depseek quickly showed how Technologies do not progress linear but exponential Already a rhythm that accelerates day by day creating a global impact not only on the progress in itself but also on money flows to global companies.
The threat of losing much of the market is the great concern, remember that USA They have limited to China the purchase of latest technology chips and only sell chips with less performance or limited. Today the most modern ones occur in Mexico for those limitations that currently manufacture the GB200 that is the most powerful until today and cannot export them to China by political decisions.
With these limitations this generative the generative of Chinese origin was created, achieving lower development and training costs by 90%matching and even improving the performance of the answers. And while technological changes always tend to lower access costs, we are talking about a Radical change in a very short period of time And that is why the market was definitely expected.
Source: Ambito
