On March 15, 2023, OpenAI made a groundbreaking announcement about the launch of ChatGPT, an advanced AI chatbot set to replace its successful predecessor from the previous year. Initially, the privilege of using the new GPT-4 language model, considered a significant leap from GPT-3.5 powering ChatGPT currently, will be limited to ChatGPT Plus subscribers and developers.
The eagerly awaited livestream event focused on the exciting potential of GPT-4’s rumored multimodal capabilities, enabling the chatbot AI to handle various input types such as text, images, and potentially even videos. While the demonstration mainly showcased text and images, it left us curious about the future possibility of diverse output formats.
During the live stream, an impressive showcase involved the AI chatbot functioning as a Discord bot, remarkably transforming hand-drawn images into fully functional websites. The potential of GPT-4 to process and output in different formats opens up fascinating prospects.
Though the primary focus of the livestream was the GPT-4 API and its applications for developers, the shown features left a lasting impression. Among the demonstrations were the replacement of existing tax software using the GPT-4 model, image processing capabilities, and even building websites with the AI functioning as a Discord bot. The OpenAI GPT-4 Developer Livestream not only showcased groundbreaking technology but also brought intriguing insights into the world of AI.
One of the most remarkable improvements in GPT-4 is its ability to process 8 times more words compared to ChatGPT, with a capacity to respond in up to 25,000 words. This expanded capability allows for more in-depth answers, even summarizing entire web pages or blog posts efficiently.
The rollout of GPT-4 will initially be available to ChatGPT Plus subscribers and developers utilizing the ChatGPT API. However, it remains uncertain when or if the free version of ChatGPT will be upgraded to GPT-4.
The livestream also revealed another groundbreaking feature, as GPT-4 can effortlessly build fully functional websites based on hand-drawn diagrams, employing JavaScript and generating additional content seamlessly.
A major leap forward in comparison to the current version of ChatGPT is GPT-4’s ability to process both text and image inputs. While the prospect of video input was hinted at by Microsoft, it has not been demonstrated yet.
The highly anticipated GPT-4 model has already been announced on OpenAI’s website as a new product. With its remarkable capability to read and comprehend up to 25,000 words, exceeding ChatGPT’s limit of 3,000 characters, GPT-4 promises a truly enhanced chatbot experience. Moreover, its multimodal features, enabling image processing alongside text comprehension, are paving the way for exciting advancements in AI. The ongoing developments in the AI landscape, as exemplified by Microsoft’s press conference, indicate that the journey to innovation in AI is an ever-evolving one.