Sam Al -Tamman admits that Openai was “on the wrong side of history” in an open source discussion

Sam TamanOpenai CEO, Made Amazing acceptance On Friday, his company “on the wrong side of history” was with regard to Amnesty International, which indicates a possible seismic transformation in the strategy while intensifying competition from China and gaining effective open models.

The explicit confession came during a response.Ask me anything“The session, a few days after the Chinese company Deepseek AI The global markets shook Through an open source R1 model that claims the performance of Openai’s systems with a break.

“Yes, we are discussing (the release of typical weights),” Taman books. “I personally think we were on the wrong side of history here and we need to discover a different open source strategy.” He pointed out that not everyone participates in Openai, his point of view, not the company’s higher priority.

The statement represents a wonderful exit from Openai is increasingly In recent years, which has caused criticism from some researchers of artificial intelligence and former allies, the most prominent of them Elon MuskWhich suits the company because of the betrayal of its original mission open source.

Sam Altman on Dibsic: “We will keep less lead.”

Altman comments come in the middle Market disturbances Catured the appearance of Depsic. The Chinese company’s claims to build models of artificial intelligence $ 5.6 million in training costs (Although the total development costs It is likely to be much higherI sent Nafidia shares dropWipe nearly $ 600 billion in the market value-the largest day to any American company in history.

“We will produce better models, but we will maintain less progress than we have done in previous years,” Altman admitted in the same Ama, to address the Deepseek effect directly.

Sam German admits a closed Openai strategy

The process of completion of Deepseek, whether its specific claims were accurate or not, the dynamics of the transformation in developing artificial intelligence. The company says it has achieved its results only using 2,000 Nvidia H800 GPUSLess chips of 10,000 chips that are usually published by major artificial intelligence laboratories.

This approach indicates that the algorithm and architectural improvement may be more than the power of raw computing. The revelation not only threatens Openai’s technical strategy, but its entire business model is based on exclusive access to huge arithmetic resources.

Open discussion Source: Innovation for safety

However, Deepseek as well National security concerns. The company stores user data on servers in China, the main mainland, as it can be subject to the government. Many American agencies have already moved to restrict their use, as NASA has become the latest to ban the application, citing “safety and privacy concerns.”

The potential Openai axis to open the source will return to its roots. The company was established as a non -profit institution in 2015 with the task of ensuring the humanitarian intelligence entitlements. However, his move to the “profit -covered” model and the increasingly closed approach has attracted criticism from defenders of the open source.

“The correct reading is:” Open sources models go beyond monopolistic models, “the head of Amnesty International wrote in Meta. Yan to be on LinkedInIn response to the appearance of Depsic. They reached new ideas and remained at the head of others. Since their work is published and an open source, everyone can benefit from it. This is the open search force and open source. “

A new chapter in developing artificial intelligence

Although Altman’s comments indicate a strategic shift that may be coming, he stressed that the open source is not the top priority of Openai currently. This frequency reflects the complex reality facing artificial intelligence leaders: achieving a balance between innovation, security and marketing in the Amnesty International multi -polar world.

The risks extend beyond the final result of Openai. The company’s decision can reshape the entire AI. The main open source models can accelerate innovation and give the democratic character access, but it may also complicate efforts to ensure the safety and security of artificial intelligence-the Openai mission principles.

The timing of Altman’s acceptance, after the Deepseek market instead, indicates that Openai may have a reaction to the market forces instead of leading them. This interactive position represents a remarkable role to look at the company that has long laid itself as a North star of artificial intelligence.

With the stability of dust from the appearance of Deepseek, there is one clear thing: the real turmoil is not only related to technology or market value – it is related to the challenge to assume that artificial intelligence models that are guarded are the most beautiful to the artificial general intelligence. In this light, the acceptance of Altman may be less than that it is on the wrong side of history and more than that of the realization that history itself has changed the path.