Prafulla Dhariwal, head of the Openai multimodal, discusses the tendency of the Ghibli, the role of India in its success and the future of creativity fueled by AI.
Dhariwal (30) admits that he was really blown away by the overwhelming response, especially India.
“We knew that we were building something significant, but the extent of the answer – in particular India – completely breathtaking us. In the weeks following the launch, India became our fastest market in the world for the Image generation. Globally, this has resulted in the greatest increase in the growth of the history of Openai,” said Dhariwal, the Multimodal chief, Interview.
The outgoing creativity of India was incredible, explains Dhariwal, who did his schooling of Dr. Kalmadi Shamrao High School in Pune. His academic activities took him to the prestigious Massachusetts Institute of Technology (MIT) from where he graduated in 2017. He joined Openai in the summer of 2016 as an intern and became a full -time employee a year later.
Dhariwal says that the journey of life is a will that the path of the school of Pune to work on the border AI is “absolutely possible” with persistence, humility and a clear meaning of the objective.
From selfies to emblematic photographs, users through India have adopted the tendency of “ghiblification”, making the GPT-4O images generator, a familiar name and marking a new controversial chapter in creativity fueled by AI.
Last year, the CEO of Openai, Sam Altman, praised Dhariwal, saying that GPT-4O would not have occurred without his “vision, talent, conviction and determination”.
The genius of Pune Tech says that the launch of the generation of images in GPT-4O was “one of the most rewarding steps” of its time in Openai, both technically ambitious and energizing in a creative way.
In India last December for his wedding preparations, Dhariwal says he used the model to create design ideas for the wedding card. Seeing something that he and his team had built, being part of such a significant moment “was really special,” he said.
“Our five-year team spent almost a year building a generation of 4O images from zero-in order to make the generation of images faster, clearer and more intuitive overall,” he said.
In fact, the operating director of Openai, Brad Lightcap, had written a post on X earlier this month by going the first week for chatgpt images as “very crazy” with more than 130 million users who generated 700 million more short-term images. India is now the fastest growth chatpt market in Openai.
Asked about concerns about questions such as copyright and ethics, Dhariwal said that the company’s objective was to give users a significant creative freedom while respecting artistic and legal borders.
“We do not allow generations in the style of individual living artists, but allow wider studio inspiration styles, which are legally authorized under copyright law-because the style itself is not protected by copyright. Ghibl-gholes outings are often original,” he said.
As the use of the real world is evolving, the same goes for business policies, he said and added “we learn and we are continuously adapting”.
For the future, Openai is “excited” as to the potential of new types of collaboration with animation studios and artists to co-create authentic and licensed experiences, he added.
Dhariwal says it was humiliating to help build something that is not only fun to use, but part of the way people create and express themselves.
“The outgoing creativity of India has been incredible. People use it to create family portraits, recreate scenes from Bollywood, design wedding invitations and even planning proposals. As a fan passionate about cricket, see Sachin Tresulkar sharing a image generated with our tool was a surreal moment-it really struck me so much. said.
Dhariwal says that he and the team did not focus on any particular style during research or training; And instead, aimed at building a flexible and capable system that could manage a wide range of visual and aesthetic prompts.
The Ghibl-Est style look that has become one of the most appreciated outings, but this was just an example of what the model could do, he underlines.
The famous technician says that his parents are incredibly proud of his work on the generation of images in GPT-4O, especially after seeing the incredible answer of India.
“I grew up in a family that cared deeply about learning and difference. So doing work on advanced AI that has the potential to reach and empower people on a large scale-it’s quite special,” he said.
Dhariwal’s message to young researchers, aspiring the scientists of AI and young people from India: stay curious, continue to learn and be oriented on the solution.
“To young researchers, in particular in India, my message is simple. Adopt a state of growth. Stay curious. Continue to learn. Do not turn away from difficult problems, even if they feel out of reach at first. Do you have to become that you can become the person who can resolve them,” he said.
The AI moves quickly, and there is room so that people from all walks of life contribute significantly, he said.
In ten years, Dhariwal hopes to have played a role in creating tools that really allow people to be able to students, teachers, designers, engineers or simply curious minds to explore new ideas. Especially in India, where there is an immense creativity and unexploited potential, the opportunity to make a real impact is incredibly significant, he says.
“If the tools we create can help someone work more efficiently, learn more intuitively or express themselves in a way that they could never before, it would be deeply fulfilling,” he said.