OpenAI CEO Sam Altman clarified his previous remarks regarding the development costs of advanced AI models during a fireside chat with India’s IT Minister Ashwini Vaishnaw on Wednesday. This marked Altman’s second visit to India, a country he described as an essential market for the AI giant. He is also schedule to meet Prime Minister Narendra Modi later in the day.
‘Taken Out Of Context’
Addressing earlier statements that sparked controversy, Altman acknowledged that his comments had been misinterpreted. “That’s a reference to a comment I made here a few years ago about cost,” he explained. “Maybe that was taken out of context. That was at a very specific time when there was a certain scaling thing, where I thought, and I still think, to stay on the frontier of pre-trained models, is expensive,” he added.
In 2023, Altman’s assertion that it was “totally hopeless” for smaller teams with limited budgets to compete with OpenAI in training foundational models had stirred significant debate within the AI community.
ALSO READ: DeepSeek Who? OpenAI’s New ‘Deep Research’ Tool Can Do The Job Of A Research Analyst In Seconds
‘Not Cheap, But Doable’
During his current visit, Altman highlighted advancements that have made AI development more accessible. He praised breakthroughs in distillation techniques that allow for the creation of smaller, more efficient models. “We learned a lot to do with small models. And these reasoning models in particular can be, it’s not cheap, it’s still expensive to train them, but it’s doable,” he noted.
Altman further elaborated on the evolving economics of AI model development. “We can look at costs in two different ways. These costs will continue to rise on an exponential curve, but the returns from increasing intelligence will also be exponential in terms of the economic value and the scientific value that you can create,” he said.
He emphasised that despite reductions in development expenses, demand for AI hardware would continue to grow. “Lower costs will lead to wider adoption, driving AI usage for a multitude of new applications,” Altman explained.
DeepSeek Impact?
His visit comes amid growing attention on DeepSeek, a Chinese foundational AI model that has demonstrated comparable results to Western models while utilising less expensive and less advanced chips.
DeepSeek’s success has prompted discussions about whether training large language models truly requires billions of dollars in capital and cutting-edge semiconductor technology. Altman himself acknowledged the significance of DeepSeek’s progress, describing its R1 model as “impressive” for the limited funding it reportedly received.
ALSO READ: DeepSeek To Be Hosted On Indian Servers, Ashwini Vaishnaw Confirms. Here’s How It Will Be Possible
‘India Incredibly Important Market’
IT Minister Vaishnaw highlighted India’s ambitions in the AI space, including efforts to design chips and develop cost-effective foundational models. India’s entrepreneurial talent is focused on achieving the next level of innovation, Vaishnaw said, drawing parallels to the country’s success with the Chandrayaan mission. “Why can’t we do the same in the LLM space?” he questioned.
Altman reiterated India’s importance to OpenAI, calling it the company’s second-largest market. “India is an incredibly important market for AI in general and OpenAI in particular. It is our second-biggest market, and we tripled our users here last year,” he revealed, underscoring the country’s growing role in shaping AI’s future.
