Following a day of product launches, researchers, engineers, and executives from OpenAI, including its CEO, Sam Altman, participated in an extensive Q&A session (AMA) on Reddit on Friday.
OpenAI finds itself in a delicate situation. The company is battling the perception that it is losing ground in the AI race to Chinese companies like DeepSeek, which it accuses of stealing its intellectual property. While trying to strengthen its ties with Washington and advancing an ambitious data center project, OpenAI is also reportedly preparing one of the largest funding rounds in history.
Altman acknowledged that DeepSeek has narrowed OpenAI’s lead in AI and admitted that the company has been “on the wrong side of history” when it comes to open-sourcing its technologies. Although OpenAI has released the code for some of its models in the past, it has generally favored a more closed and proprietary approach.
“I think we need a different strategy regarding open source,” Altman stated. “Not everyone at OpenAI shares this view, and it’s not our top priority right now… We will develop better models in the future, but our lead will be smaller than in previous years.”
Kevin Weil, OpenAI’s product director, mentioned that the company is considering open-sourcing older models that are no longer state-of-the-art. “We are definitely thinking about doing more things like that,” he said, though he did not provide further details.
Altman also noted that DeepSeek has pressured OpenAI to reveal more about how its reasoning models, such as the recently launched o3-mini, demonstrate their “thought process.” Currently, OpenAI’s models hide this reasoning to prevent competitors from extracting training data. In contrast, DeepSeek’s R1 model displays its entire chain of thought.
“We are working to show much more than we do today; [showing the thought process] will come soon,” Weil added. “It’s still unclear how this will be handled; showing the entire chain of thought could lead to a competitive advantage, but we know advanced users want it, so we’ll find the right balance.”
Altman and Weil dismissed rumors that ChatGPT, OpenAI’s chatbot platform, would increase its prices in the future. In fact, Altman expressed his desire to make ChatGPT “cheaper” over time, if possible.
Previously, Altman had admitted that OpenAI was losing money on its ChatGPT Pro plan, which costs $200 per month.
Weil also emphasized that OpenAI continues to observe that greater computing power leads to more powerful and higher-performing models. This is what drives projects like Stargate, the recently announced massive data center project. Additionally, the growing user base is increasing the demand for computational resources within the company.
Regarding the possibility of these powerful models enabling recursive self-improvement (a process where AI enhances its own intelligence without human intervention), Altman commented that a “fast takeoff” is more plausible than he once believed.
It’s worth noting that Altman is known for making ambitious promises. Not long ago, he tempered OpenAI’s expectations regarding artificial general intelligence.
A Reddit user asked whether OpenAI’s models, with or without self-improvement capabilities, could be used to develop destructive weapons, specifically nuclear weapons. Recently, OpenAI announced a collaboration with the U.S. government to provide its models to National Laboratories, partly for research related to nuclear defense.
Weil expressed his confidence in the government: “I know these scientists; they are AI experts and world-class researchers. They understand the limits and potential of the models, and I don’t think they would simply turn a model’s output into nuclear calculations. They are meticulous, evidence-based, and conduct extensive testing to validate their work.”
The OpenAI team also answered technical questions, such as when the next reasoning model, o3, will be launched (“in a few weeks or months,” said Altman), when GPT-5 will arrive (“there’s no timeline yet,” Altman admitted), and when the successor to DALL-E 3, OpenAI’s image generation model, will be revealed. DALL-E 3, launched two years ago, has become outdated compared to recent advances in image generation technology.
“Yes, we are working on it,” Weil said about the successor to DALL-E 3. “And I think it will be worth the wait.”
By: Nestor Castillo, ForAllTechNews Director