OpenAI has unveiled GPT-4 Turbo, offering responses with knowledge up to April 2023 – a marked improvement over its predecessor’s January 2022 cutoff date. The context window (the amount of text you can use in your prompts and conversations) has been increased from about 3,000 words to approximately 300 pages, making the platform significantly more useful for real world tasks.
The upgrade promises to remove “feature selection” from ChatGPT’s pulldown menu, so you will no longer have to guess whether you want to “Browse with Bing” or use a plug-in or use the default model; ChatGPT Plus will “just know.”
OpenAI also announced that ChatGPT users will be able to create custom “GPTs,” which are customized, special knowledge versions of ChatGPT. (You can think of them as personalized chatbots.) There’s a GPT store on the way, too, so you’ll be able to share and (I assume) profit from your specialized creations.
Interestingly, one of the biggest announcements didn’t draw a sound from the otherwise enthusiastic audience. As almost a throw-away, OpenAI’s CEO Sam Altman promised that the company will defend its customers against copyright infringement claims, removing the most significant adoption barrier for corporate clients. (Microsoft, Google, Adobe, and other big AI providers already offer this indemnification.)
This was a very big day for OpenAI and for everyone thinking about using generative AI platforms to increase productivity. You can find a full list of OpenAI’s announcements here.
Author’s note: This is not a sponsored post. I am the author of this article and it expresses my own opinions. I am not, nor is my company, receiving compensation for it. This work was created with the assistance of various AI models, including but not limited to: GPT-4, Bard, Claude, Midjourney, Stable Diffusion, and others.