ChatGPT Tests New Feature: Automatic Tool Selection Based on Context

October 30, 2023 – OpenAI is rolling out new beta features for ChatGPT Plus members. Users have reported that this update includes the ability to upload and process files, as well as support for multimodal capabilities. This means that users no longer need to select modes like “Browse with Bing” from the GPT-4 dropdown menu. Instead, the model automatically guesses what users want to do based on the context.

These new features bring ChatGPT Plus members some of the office functionalities offered by the ChatGPT Enterprise plan. It’s been reported that after uploading a file to ChatGPT, it takes a few seconds for the model to digest the file, and then it can perform tasks such as summarizing data based on prompts, answering questions, or generating data visualizations.

It’s worth noting that the new ChatGPT features are not limited to processing text files. On Threads, a user shared a screenshot of a conversation in which they uploaded an image of a sea otter. ChatGPT, using DALL-E 3, created a Pixar-style image based on the sea otter. The user then uploaded another image, this time of a distorted skateboard, and asked ChatGPT to insert it into the previously generated image. ChatGPT successfully accomplished this task, even adding a hat to the sea otter, although the reason for the hat remains a mystery.

Leave a Reply