During OpenAI’s inaugural developer conference, CEO Sam Altman shared that ChatGPT, the widely popular AI chatbot, has achieved a milestone of 100 million weekly active users. This was coupled with the introduction of the latest breakthrough, GPT-4 Turbo, an enhanced and more cost-effective iteration of OpenAI’s flagship text-generating AI model, GPT-4.
Since its launch merely a year ago, ChatGPT rapidly gained traction, amassing an estimated 100 million monthly users within the initial two months. Altman underscored that the platform is now leveraged by over two million developers, including 92% of Fortune 500 companies.
A key highlight of the conference was the unveiling of GPT-4 Turbo, the next-gen model equipped with a 128k context window, enabling it to comprehend the equivalent of more than 300 pages of text in a single prompt. This advanced version not only boasts increased capabilities but also offers a more cost-effective solution, with input token prices reduced by a third and output token prices by half compared to GPT-4. GPT-4 Turbo, equipped with knowledge up to April 2023, is currently available for preview to paying developers and is scheduled for release in the upcoming weeks.
In addition, OpenAI introduced the Assistants API, streamlining the process for developers to create assistive AI applications with specific objectives. The API incorporates features such as Code Interpreter, Retrieval, and function calling, with enhancements in function calling accuracy, improved instruction following, and the introduction of a JSON mode for precise output generation.
Expanding the capabilities of GPT-4 Turbo, OpenAI integrated vision support, enabling it to process images and undertake tasks such as generating captions and analyzing real-world images. The integration of DALL•E 3, a technology allowing the programmable generation of images and designs, further accentuated the capabilities of the platform. OpenAI also introduced a text-to-speech API, facilitating the production of high-quality speech from text, offering users a selection of six preset voices.
To provide further support to developers, OpenAI introduced an experimental access program for GPT-4 fine-tuning, catering to a tailored experience for specific domains. Furthermore, the Custom Models program was announced, permitting select organizations to collaborate closely with OpenAI researchers to create highly customized solutions.
The event was marked by a surprise appearance from Microsoft’s CEO, Satya Nadella, emphasizing the partnership between the two tech giants. Nadella commended OpenAI’s commitment to safety and empowerment, with Altman emphasizing the importance of a cautious and gradual approach to AI deployment.
Leave a comment