After some mega cap leaders expressed fears for the development of artificial intelligence recommending one slowdown and causing the market crash, now we go in search of the antibodies able to give a more relaxed reading. From Zuckerberg of Meta, to Jensen Huang of Nvidia it is now emphasized that a false alternative between rapid innovation and AI safety and that can be found the right measures inside. Meanwhile the megacap of AI are demonstrating in fact who have not no intention of slowing down and continue on their way, taking care of the upcoming IPOs, collecting funding, creating new data centers.
Zuckerberg presents Meta as a responsible actor on AI safety and suggests independent evaluators
The CEO of Meta Platforms, Mark Zuckerberg seems to want to distance yourself from the appeals of the leaders of the major AI companies stressing that there are sufficient reasons for act individually in the field of of your digital ecosystem. . In a post on X He stressed that each laboratory has the responsibility and incentive to proceed at the pace necessary to train its own models in of your digital ecosystem. , and the ability to take action to ensure that this happens, possibly by relying on independent evaluators and consultants for ensure safety.
His comments come three days after Anthropic's CEO, Dario Amodei, He urged AI companies to slow the pace at which they improve the capabilities of their models, following warnings of existential risks from several researchers. Sam Altman, CEO of the rival OpenAI, and Elon Musk of xAI both publicly supported Amodei's appeal within hours.
Zuckerberg also said that labs have a strong incentive to prevent their models from causing harm, as they would otherwise be at risk of face “significant” responsibilities. He cited as an example the delay in releasing Meta's AI agent Muse earlier this year to strengthen its security. Meta had already allocated "most" of its computing power to creating products that meet immediate user needs. instead of pursuing self-improving AI systems. Zuckerberg also stressed that Meta Superintelligence Labs, Its artificial intelligence division already employs independent evaluators in several areas, calling it "best practice in the industry" and adding that "other labs can do the same." The company has agreed to pay up to $18 billion to settle lawsuits filed by U.S. states accusing it of designing Facebook and Instagram to addict children, but has not admitted any wrongdoing.
Nvidia CEO: Panic has created a false choice between rapid innovation and AI safety, but it's absolutely possible to have both
Yesterday the president Donald Trump He downplayed concerns about the potential harms of artificial intelligence, saying that the United States already has safeguards in place for the technology and that it would be the China to take advantage of the doubts that are being raised about the development of AI. Trump also expressed his concept by intervening with a live call al Nvidia CEO Jensen Huang on stage at the All-In Summit in Los Angeles. Trump criticized the push to slow the pace of AI development, calling it politically motivated or a Chinese conspiracy. Huang, while agreeing with the president in some respects, also downplayed the risks of AI, adding that panic has created a false alternative between rapid innovation and AI safety, but it is absolutely possible have both things at the same time.” Only at the end of August did Huang predict that the surge in sales driven by artificial intelligence will last until 2028.
No AI company has any intention of slowing down. On the contrary.
Meanwhile the megacap continue to go their own way as they had planned. Moreover, none of the promises made by the major operators in the sector, or by the main AI companies, mention a possible reduction in the infrastructure development. On the contrary. Precisely in the essay by Dario Amodei, which has caused such a stir, it is emphasized that “slowing down doesn't mean stopping model training or technical progress.”
At the moment of greatest panic on the markets, on the idea of a slowdown in the development of AI, the statement of OpenAI to postpone his listing next year. In reality the roadmap it was already set up this way since this summer, and the creator of ChatGPT has not shelved the project due to the latest statements. So much so that in these hours OpenAI itself has started preliminary discussions with investors for a new round of financing which would be valued at over 1.200 trillion dollars.
The anthropic Anthropic is proceeding as planned and, after raising funds in May at a valuation of $965 billion for its IPO, has chosen the Nasdaq as its listing market. The company that produces the Claude chatbot aims to raise a sum equal to or greater than SpaceX's in an IPO that could take place as early as October, according to Bloomberg. SpaceX raised a record $86,3 billion in its initial public offering in June. Anthropic itself also just signed its first deal for a data centers in Australia, back Reuters, with a planned capacity of 2,16 gigawatts, comparable to some of the largest centers currently operating in the world. The proposed facility, located on agricultural land approximately 250 km from Brisbane, is expected to begin operations in 2027.
Same MetaPlatforms announced that he wanted to proceed at full speed with the installation of its own AI chips in data centers by early next year. This from a company that claimed to have suffered a cyberattack by a rogue AI in August, much like OpenAI's Hugging Face fiasco. Finally Elon Musk, with a comment to be taken with a grain of salt, he outlined his plans for to bring into space the AI platform Nvidia's Vera Rubin next year.
