Openai CEO Altman itself revealed that his company reached 800 million weekly energetic users and has “unbelievable” growth rates, during a time tense conversation in Ted 2025 Conference at Vancouver last week.
“I have never seen a growth in any company I was involved with or not, in this way,” he told Altman to the boss Chris Anderson during a conversation on stage. “Development of chatgpt – it’s really cool. I feel deeply honored. But it’s crazy to experience and our teams are exhausted and stressed.”
An interview that ended the last day Ted 2025: Humanity is again imaginedHe presented not only the rapid success of OpenAI, but also the growing control, which is the company, when its technology transforms society at a pace, which even alerts some of its supporters.
“Our GPUs are melting”: Openai is fighting for scaling among unprecedented demand
Altman painted a photo of a company fighting for maintaining with its own success, noting that the OPENAI GPU “melts” on account of the popularity of latest image generation functions. “I call people all day and begging them to give us their GPU. We are so extremely limited,” he said.
This exponential growth appears when OpenAi is supposedly considering Launching your personal social network According to CNBC, Elon Musk, based on CNBC. Altman has not confirmed or denied these reports during Ted’s interview.
The company has recently closed Financing round $ 40 billionPriceing it at $ 300 billion – the largest private technology funds in history – and this influx of capital will probably help to deal with some of those infrastructure challenges.
From non-profit organization as much as $ 300 billion: Altman responds to the accusations of “Ring of Power”
During the 47-minute conversation, Anderson repeatedly pressed the Altman to the OpenAI transformation from the non-profit research laboratory in a company focused on profit with a valuation value $ 300 billion. Anderson expressed concerns shared by critics, including Elon Musk, who suggested that Altman was “broken by the ring of power”, referring to the “Lord of the Rings.”
Altman defended the OpenAI path: “Our goal is to make Aga and distribution to it, ensure its wide good thing about humanity. I think that we have done a lot in all accounts in this direction. Apparently our tactics have modified over time … We didn’t think that we’d have to build a company around this company. We learned a lot about the way it goes and the realities of those systems.
When asked how he personally deals with the huge power, which he runs now, Altman replied: “Shockingly, as before. I think you can get used to everything step by step … You are the same person. I am sure that I am not in different ways, but I do not feel different.”
“Division of revenues”: Openai plans to pay to artists whose styles are used by AI
One of the most specific announcements of the policy from the interview was confirmation of Altman that OpenAI is working on the system of compensating artists whose styles are imitated by artificial intelligence.
“I think there are amazing new business models that we and others are excited about discovering,” said Altman when we press the apparent IP theft in paintings generated by AI. “If you say:” I need to generate art in the sort of those seven people, everyone has agreed to it: “How do you share, how much money does it go to everyone?”
Currently, the OpenAI image generator rejects requests to mimic the sort of living artists without consent, but generates art in the sort of movements, genres or studies. Altman suggested that the model of division of revenues could be available, although the details remain rare.
Autonomous AI agents: Openai’s “most consistent security challenge”
The conversation became particularly tense during the discussion “Agentic ai” – Autonomous systems that may take motion on the Internet on behalf of the user. New” OpenAI “Operator“The tool allows artificial intelligence to perform tasks reminiscent of restaurant booking, raising concerns about security and responsibility.
Anderson challenged Altman: “One person could allow this agent, and the agent could decide:” Well, to perform this function, I have to repeat in every single place. ” Are there red lines that you clearly drew internally, where you know what dangerous moments are? “
Altman mentioned “Openai”Readiness framework“But he provided little details about how the company would prevent the incorrect use of autonomous agents.
“Ai that you give access to your systems, information, the ability to click the computer … When they make a mistake, it’s much higher rates,” confirmed Altman. “You will not use our agents if you do not trust that they do not intend to empty the bank account or delete data.”
’14 Definitions from 10 researchers: inside the Openai fight for define AGA
At an revealing moment, Altman admitted that even at Opennai there is no consensus as to what is an artificial general interview (Agi) – the set goal of the company.
“It’s like a joke, if you have 10 OPENAI researchers in the room and ask to define Aga, you will receive 14 definitions,” said Altman.
He suggested that as an alternative of focusing at a specific moment when Agi arrives, we must always recognize that “the models will be simply smarter, more capable and smarter and more capable of this long exponential … We will have to fight and get great benefits from this amazing system.”
Loosening the handrail: OpenAi latest approach to content moderation
Altman also revealed a significant change in content moderation rules, revealing that Opeli loosened restrictions on image generation models.
“We gave users much greater freedom in what we traditionally thought as damage to speech,” he explained. “I think that part of the model leveling is to follow what he wants to make the model of the model do it within the very wide limits of what society decides.”
This change can signal a broader traffic towards providing users with greater control over AI outputs, potentially adapting to the expressed Altman’s preference to enable lots of of tens of millions of users – as an alternative of “small elite peaks” – determining the appropriate handrails.
“One of the cool new things in artificial intelligence is that our artificial intelligence can talk to everyone on earth and we can learn the preferences of the collective value of what everyone wants, instead of having a group of people who are blessed by society to sit in peace and make these decisions,” said Altman.
“My child will never be smarter than AI”: Altman’s vision about the future driven AI
The interview ended when Altman thought about the world that his newborn son would inherit – one in which artificial intelligence will exceed human intelligence.
“My child will never be smarter than AI. They will never grow up in a world where products and services are not incredibly intelligent, extremely capable,” he said. “It will be a world of incredible abundance of material … where the change rate is shared, they share incredibly fast and amazing new things.”
Anderson closed the sobering commentary: “In the next few years you will have one of the greatest possibilities, the greatest moral challenges, the greatest decisions about making every person in history.”
The act of balancing billions of users: how OpenAi moves power, profit and purpose
Altman’s appearance is critical for OpenAI and wider AI industry. The company faces growing legal challenges, including Copyright From the authors and publishers, while crossing the boundaries of what AI can do.
Last progress, reminiscent of Chatgpt’s Viral image generation function And the Sora video generation tool showed opportunities that seemed inconceivable only a few months ago. At the same time, these tools caused Copyright debatesauthenticity and way forward for creative work.
Altman’s readiness for difficult questions about the safety, ethics and impact of social and the social of the social rates shows the awareness of the rates involved. However, critics may notice that specific answers about specific security and rules remained elusive during the conversation.
The interview also revealed competitive tensions at the Openai Mission Center: moving quickly to develop AI technology while ensuring safety; Balancing the motives of profits with social advantages; respect for creative rights in the democratization of creative tools; and navigation between elite knowledge and public preference.
As Anderson noted in his final commentary, the decisions are made by Altman and his peers in the coming years, may have an unprecedented influence on the way forward for humanity. Whether Openai can fulfill its mission to make sure “all of humanity from artificial general intelligence” will only prove.