Join our daily and weekly newsletters for the latest updates and the exclusive content on AI coverage. Learn more
Sam AltmanCEO of Openai, made a Striking admission Friday, that his business was “on the wrong side of history” concerning open source AI, reporting a potential seismic change in strategy because China competition intensifies and that effective open models are gaining ground.
Candid recognition came during a reddit “Ask me”Session, a few days after the Chinese IA company Deepseek World markets shaken With its Open Source R1 model which requires performance comparable to Openai systems at a cost fraction.
“Yes, we discuss [releasing model weights]”” Altman wrote. “I personally think that we have been on the wrong side of the story here and that we must find a different open source strategy.” He noted that everyone in Openai does not share their point of view and that it is not the current priority of the company.
The declaration represents a remarkable gap Increasingly owner approach of Openai In recent years, which has aroused criticism from certain researchers and former allies, in particular Elon MuskWho continues the company for having allegedly betrayed its original open source mission.
Sam Altman on Deepseek: “ We will maintain less an advance ”
Altman’s comments come in the middle Turmoble walking triggered by the emergence of Deepseek. The allegations of the Chinese company for the construction of advanced AI models for $ 5.6 million in training costs (although total development costs are Probably much higher) sent the Nvidia stock fallAnnouncement nearly $ 600 billion in market value – the largest decrease of one day for any American business in history.
“We will produce better models, but we will maintain less an advance than we have done in previous years,” said Altman in the same AMA, directly approaching the impact of Deepseek.
Sam Altman admits that OpenAi’s closed strategy can be defective
The breakthrough of Deepseek, whether his specific affirmations are accurate or not, has highlighted the dynamics of change in the development of AI. The company says it has obtained its results using only 2,000 GPU NVIDIA H800—Far less than the more than 10,000 chips estimated generally deployed by the main AI laboratories.
This approach suggests that algorithmic innovation and architectural optimization could have more importance than raw calculation power. The revelation does not only threaten OpenAi’s technical strategy, but its entire commercial model has occurred on exclusive access to massive IT resources.
Open source debate: Innovation vs security
However, the rise of Deepseek has also intensified National security concerns. The company stores user data on servers in continental China, where it could be subject to government access. Several American agencies have already decided to restrict its use, NASA becoming the last to block the application citing “security and confidentiality problems”.
Openai’s potential pivot to Open Source would mark a return to its roots. The company was founded as a non -profit organization in 2015 With the mission to guarantee that artificial general intelligence benefits humanity. However, its transition to a “capped” model and an increasingly closed approach have aroused criticism from open source defenders.
“Correct reading is:” Open Source models exceed the owners “, wrote the main scientist of Meta AI Yann Lecun on Linkedinresponding to the emergence of Deepseek. “They proposed new ideas and built them above the work of others. Because their work is published and open source, everyone can take advantage of it. It is the power of open research and open source. »»
A new chapter in the development of AI
Although Altman’s comments suggest that a strategic change can happen, he stressed that open source is not currently the absolute priority of Openai. This hesitation reflects the complex reality faced with AI managers: balance innovation, security and marketing in an increasingly multipolar AI world.
The stakes extend far beyond the OPENAI net profit. The company’s decision could reshape the entire IA ecosystem. The key models of open source could accelerate innovation and democratize access, but this could also complicate efforts to guarantee the security and security of AI – the principles of the Openai mission.
The time of altman’s admission, coming after the Deepseek market shock rather than before, suggests that Optaai could react to market forces rather than directing them. This reactive position marks a striking role reversal for a company that has long positioned itself as North Star of AI.
While the dust sets in the beginnings of Deepseek, one thing becomes clear: the real disturbance does not only concern technology or market value – it is a question of challenging the hypothesis that the closely kept AI models are the The safest way to general artificial intelligence. With this in mind, Altman’s admission could be less concerning the bad side of history and more on recognition than the story itself has changed CAP.
