In a wide-ranging interview with MIT researcher and popular podcaster Lex Fridman, OpenAI CEO Sam Altman shared insights on the company's latest innovations and his vision for the future of artificial intelligence. The conversation touched on OpenAI's new multimodal AI system Sora, the development of GPT-5, and Altman's perspectives on the path to artificial general intelligence (AGI).
Altman expressed excitement about Sora's ability to understand and generate images and video, seeing it as an important step towards AI systems that can effectively operate in the physical world. Despite current limitations, he expressed confidence that continued scaling and refinement will rapidly advance Sora's capabilities.
Looking ahead to GPT-5, Altman emphasized the significance of the model becoming "smarter" in a generalized sense rather than just improving at narrow tasks. "I expect that the delta between 5 and 4 will be the same as between 4 and 3," he stated. When asked about a release timeline, Altman avoided specifics but said OpenAI plans to share "an amazing new model this year."
The interview also delved into last year's dramatic shakeup of the OpenAI board, which saw Altman briefly removed as CEO. He characterized the experience as "the most painful professional experience of my life" but expressed gratitude for the support he received and his determination to keep moving the company's mission forward.
Altman also addressed the lawsuit recently filed against OpenAI by Elon Musk, dismissing Musk's claim that the company has strayed from its original "open source" principles. He argued the suit does not seem legally serious and wished the visionary entrepreneur would focus on "building cool shit" rather than attacking OpenAI.
On his relationship with OpenAI cofounder and former board member Ilya Sutskever, who played a central role in the board drama, Altman said "I love Ilya. I have tremendous respect for Ilya," but declined to reveal anything more on Sutskever's future involvement with the company.
The conversation also delved into Altman's views on AGI. He argued the term is poorly defined and prefers focusing on the development of highly capable systems that can meaningfully accelerate scientific progress and reshape the global economy. By the end of the decade, Altman predicts we will have AI systems that evoke a sense of awe in their abilities.
Other topics discussed include the need for massive increases in compute to power future AI systems, the importance of AI safety and alignment, the competitive landscape in AI development, and OpenAI's own technical approach and guiding philosophies.
Throughout the interview, Altman offered frank and thought-provoking commentary on some of the most consequential issues surrounding artificial intelligence, shedding light on some of OpenAI's tumultuous recent history and current challenges, all while providing a glimpse of the astounding AI breakthroughs that may lie ahead. If you have a couple hours, it is well worth the watch!