OpenAI’s ChatGPT is not done adding new features just yet. The highly popular chatbot is getting new and useful functionalities in its latest beta as reported by many paid users.
The AI tool now allows users to upload their own files for analysis, enabling them to ask questions about the uploaded files. This feature is limited to ChatGPT Plus, which is the ChatGPT’s paid tier that also grants you access to GPT 4 and quicker response times.
The feature is called Advanced Data Analysis and it is essentially a glimpse of the features offered with ChatGPT Enterprise, which is equipped with business-grade features.
Even though the feature is still in beta, it’s still working as intended. All you have to do is select a file to upload and then ChatGPT will take a few moments to read it. Once it’s done reading, it is ready to work on the document. It can summarize the files, answer questions, or create data visualizations from prompts.
For instance, if a user happens to have a comprehensive financial report on their hands, they can have ChatGPT help them get through it much faster. They can get explanations for lengthy datasets and also pinpoint the most important figures in the financial data.
Also Works with Images
Better yet, the new feature is not limited to files. It also allows users to use it for images and get creative with it, as noted by a user on Meta’s Threads. User @luokai posted screenshots of how he was able to post an image to ChatGPT and modify it.
The user uploaded an image of a capybara and told ChatGPT to turn it into a Pixar-style animated image. As shown in the image above, the chatbot was able to produce a very convincing result.
Then the user asked ChatGPT to add an additional element to the image it created. He asked the AI to add a wiggly skateboard to the photo. The result was not exactly as requested, but still rather impressive. For some reason, it gave the capybara a hat too.
As mentioned earlier, the new feature is only available in beta for now and it will be a while before it rolls out far and wide. That being said, multimodal support is also not available for some users in ChatGPT beta as of yet.
‘Chat with Documents’ wrappers
OpenAI is just rolling this out but there are already several products that leverage OpenAI’s language models, enabling users to interact with their documents and analyze them. One of the leading products in this space is PDF.ai, which charges users $15 per month for unlimited document analysis. In comparison, ChatGPT’s subscription costs $20 per month and offers significantly more capabilities.
Damon Chen, the founder of PDF.ai, responded to the news on X, expressing his optimism about the future of his PDF chatbot built on top of OpenAI’s technology. He stated, “I don’t think ChatGPT will ever implement small PDF-related features that customers desperately ask for. Sure, I believe small players will go away or not even want to get started, and big players with VC money will die once they burn all their money soon. PDF.ai is bootstrapped and profitable with a very healthy margin. We don’t have a mission to become another unicorn; several million dollars in ARR is good enough for me, and that’s my goal in the next 1-2 years.”