OpenAI, the renowned artificial intelligence (AI) developer, has discreetly updated the training dataset for its cutting-edge model, GPT-4 Turbo, with the changes effective from December 2023, as disclosed in documents on its website. This latest update positions Turbo as the most current model available from OpenAI, contrasting with its predecessor, GPT-3.5, which only incorporates data up to January 2022.
The motivation behind this update, as outlined on the website, is to address what developers refer to as "inertia" within the model, which pertains to its prior limitations in completing tasks. Notably, OpenAI had previously indicated at November's DevDay that GPT-4 Turbo would be cognizant of events until April 2023. Consequently, the recent update furnishes the model with an additional eight months' worth of data and insights.
However, the rollout of this update did not receive widespread announcement, and some users raised questions about its legitimacy on the OpenAI developer forum following its implementation on January 25. Despite the website reflecting updated data as of December 2023, certain users observed that their previews did not reflect this recent timeframe, instead returning test prompts referencing April 2023, which was the previous update window.
This latest development in OpenAI's technological arsenal comes on the heels of the launch of Sora, an advanced text-to-video model that has garnered significant attention on social media for its remarkable real-world output. Sora boasts the capability to generate movie-like scenes featuring multiple characters, specific actions, themes, and intricate background details, all at up to 1080p resolution. However, OpenAI has clarified that Sora is not yet ready for public release.
In a separate report on February 17, the New York Times highlighted OpenAI's staggering valuation, which has surged to $80 billion in its latest deal. The company is presently engaged in discussions with investors worldwide, as well as various international governments, with the aim of securing funds to advance its in-house development of AI chips.
