In an interview with the Financial Times published On Monday, Altman said the collaboration between OpenAI and Microsoft was “working very well,” and he expects the companies to continue investing in each other’s success.
Earlier this year, Microsoft invested $10 billion in OpenAI, valuing the San Francisco-based startup at $29 billion. Without disclosing specific financial details, Altman indicated that OpenAI experienced strong revenue growth in 2022, but remains unprofitable due to high research and development expenses associated with developing increasingly sophisticated AI models.
Deepening ties with Microsoft aims to ensure both companies can capitalize on advances in artificial intelligence technology.
Last week, OpenAI unveiled new development tools and upgrades to its GPT-4 model, which allows customization for specific applications. It also introduced a marketplace, dubbed GPT Store, where creators can monetize AI apps and share revenue with OpenAI, similar to Apple’s App Store model.
Looking ahead, Altman said OpenAI is focused on developing more autonomous software agents capable of performing increasingly complex tasks, with the goal of achieving artificial general intelligence that rivals the capabilities human.
The company is already working on GPT-5, the successor to its groundbreaking large language model, although Altman did not provide a timetable for its release. He acknowledged that it is technically difficult to predict exactly what new skills and capabilities each new iteration of technology may possess.
Like many AI researchers, OpenAI relies heavily on Nvidia’s advanced semiconductors to train its models. Altman said supply shortages of Nvidia’s expensive chips have created a “brutal crisis” in 2022, but he expects availability to improve next year as new offerings from others chip manufacturers are entering the market.
Even as OpenAI cements its lead in generative AI, Altman said the company remains focused on the elusive goal of developing artificial general intelligence. He believes that mastering natural language processing through models such as GPT is a crucial step towards this end.
Featured image credit: Photo by Karolina Grabowska; Pixels; THANKS!