All you need to know about ChatGPT's new code interpreter feature

ChatGPT

The artificial intelligence-powered Chatbot by

OpenAI

In recent months, has amazed the world with the text that it can produce. The chatbot has now delighted users by creating maps, charts, and videos from images.

OpenAI launched a new feature for ChatGPT Plus users last week called

code interpreter

What you need to know.

What is a code interpreter?

Code interpreter allows ChatGPT, among other things, to analyze data, create graphs, solve math problems, and edit files. ChatGPT now supports the uploading and download of files. Subscribers to ChatGPT Plus ($20 a month) can now access a code interpreter. Subscribers can access similar add-ons, such as Expedia or OpenTable.

What is the code interpreter?

ChatGPT guesses the answer to a question based on the large language model technology, which predicts what word will be next in a series.

OpenAI stated that when code interpreter was enabled, ChatGPT would write and run a computer program to find an answer. This allows the chatbot to perform new tasks it couldn't before, like performing complex calculations and creating charts based off data uploaded by a user, all of which are completed using code. Code interpreter, according to some, reduces the likelihood of errors. This is a common complaint with large language models. Ethan Mollick is an associate professor from the University of Pennsylvania who has tested code interpreter over a period of two months.

What has code interpreter been used for?

Data analysis is the most common application of code interpreter. ChatGPT, with a simple prompt such as "tell me what's interesting about the data", can analyze a user’s data (such as their financial, health, or location data) and provide insights. Financial analysts use the tool to analyze stock prices and prepare budgets. Researchers are also using it to create creative data visualizations. Mollick revealed that he used code interpreter recently to make a 3D chart for the Billboard Hot 100 and an animated map showing every lighthouse within the United States. Code interpreter has also been used by some people to convert file formats, such as converting images into videos and PDF documents into photos.

Arvind

Narayanan is a professor at Princeton University of Computer Science. He warned that people shouldn't rely too much on code interpreters for data analysis, as AI can still produce inaccurate results.

He said that "Appropriate Data Analysis requires a lot more critical thinking than just the data."