June 6, 2023 – OpenAI, the AI giant, has recently shared some insights during a private gathering with developers. CEO Sam Altman revealed that, apart from the already released ChatGPT, the company has no plans to launch any products that would compete with developers.
During his visit to London in May this year, Altman engaged in confidential discussions with a small group of developers and entrepreneurs. He presented OpenAI’s roadmap and attentively listened to their concerns and requirements regarding artificial intelligence. Although the meeting was intended to be confidential, one participant, Raza Habib, the co-founder and CEO of AI startup Humanloop, publicly disclosed details in a now-deleted blog post. Fortunately, the post was captured and saved by internet users.
According to the blog post, Altman expressed that OpenAI’s focus remains on providing developers with powerful language models. The company assures developers that they need not worry, as ChatGPT’s vision is to become an exceptional AI assistant for work-related tasks.
In 2020, OpenAI released its large-scale language models, available for developers to purchase and integrate into their applications. These language models enable developers to construct various applications more efficiently, minimizing the need for extensive time, expertise, and computational resources. Currently, hundreds of applications utilize OpenAI’s language models.
However, Altman acknowledged that OpenAI faces some challenges, with the most pressing being the shortage of GPUs. This limitation hinders certain short-term plans of OpenAI, such as improving the reliability and speed of their API. Altman outlined three major restrictions caused by the GPU shortage. Firstly, text length is constrained; secondly, fine-tuning processes are limited by computational power. Lastly, OpenAI struggles to find enough GPUs to offer private versions of the GPT model to their customers. If clients desire dedicated capacity, they are currently required to prepay $100,000.