To crown a day of product versions, researchers, engineers and Openai managers, including the CEO of Openai, Sam Altman, answered questions in a Reddit Ama in Vast Friday.
OPENAI, the company is in a precarious position. It fights against the perception which it gives in land in the race for AI Chinese companies like DeepseekWhich, according to Openai, may have stolen his intellectual property. The chatpt manufacturer tried to Look at his relationship with Washington and simultaneously continue a Ambitious data center projectWhile it would have dive For one of the largest financing laps in history.
Altman admitted that Deepseek had reduced the example of Openai in AI, and he also said that he thought that Openai had been “on the wrong side of history” with regard to the opening of his technologies . While Openai has open source models in the past, the company has generally favored a proprietary -owned source development approach.
“”[I personally think we need to] Discover a different open source strategy, “said Altman. “Everyone in Openai does not share this point of view, and it is not our current priority either […] We will produce better models [going forward]But we will maintain less an advance than we have done in previous years. »»
In a follow -up response, Kevin Weil, OPENAI product director, said that Optai was considering older Open models that are no longer at the cutting edge of technology. “We will certainly think of doing more,” he said, without going in more detail.
Beyond the incentive to Openai to reconsider his philosophy of liberation, Altman said that Deepseek had prompted the company to potentially reveal more about the way in which its so-called models of reasoning, like the O3-minini model published Today, show their “reflection process”. Currently, OpenAi models hide their reasoning, a strategy intended to prevent competitors from scratching training data for their own models. On the other hand, the Deepseek reasoning model, R1, shows its complete chain of thought.
“We work to show a heap more than we show it today – [showing the model thought process] will be very soon, ”added Weil. “TBD on all – showing the whole chain of thought leads to competitive distillation, but we also know people (at least power users), we want it, we will find the right way to balance it.”
Altman and Weil tried to dispel the rumors that Chatgpt, the Chatbot application through which Openai launches many of its models, would increase the price. Altman said he would like to make the Chatppt “cheaper” over time, if possible.
Altman previously said that Optai loses money on its most expensive Chatgpt plan, Chatgpt Pro, which costs $ 200 per month.
In a somewhat linked thread, Weil said that Optai continues to see evidence that more computing power leads to “better” and more efficient models. This is largely what requires projects such as Stargate, Openai recently announced Massive data center project, Weil said. The service of a growing user base also supplies the request for calculation within Openai, it also continued.
Asked about recursive self-improvement which could be activated by these powerful models, Altman said that he thought that “rapid take-off” is more plausible than it thought in the past. Recursive self-improvement is a process where an AI system could improve its own intelligence and capacities without human contribution.
Of course, it should be noted that Altman is known to promise too much. Not long ago that he Basinai bar lowered for act.
A Reddit user asked whether OPENAI models, improving or not, would be used to develop destructive weapons – in particular nuclear weapons. This week, Openai announced a partnership with the US government to give its models to American national laboratories in part to nuclear defense research.
Weil said he had trusted the US government.
“I got to know these scientists and they are IA experts in addition to world class researchers,” he said. “They understand the power and the limits of the models, and I do not think that there is a chance that they simply go a certain outing of the model in a nuclear calculation. They are intelligent and based on evidence and they have a lot of experimentation and data to validate all their work. »»
The Openai team has been asked several questions about a more technical nature, such as when the next OpenAi reasoning model, O3, will be published (“more than a few weeks, less than a few months,” said Altman), when the Next flagship product of the company “non-seasoning model, GPT-5, could land (” not having a calendar yet, “said Altman), and when Openai could reveal a successor to Dall-E 3, the generating model of ‘Image of the company. model is More competitive on a number of reference tests.
“Yes! We work there,” said Weil about Dall-E 3 follow-up. “And I think it will be worth waiting.”
