OpenAI Launches Code Interpreter for ChatGPT Plus Users
- By Paul Mah
- July 12, 2023
For those who don’t know it yet, ChatGPT has released a Code interpreter feature on ChatGPT, which it bills as an experimental model that can execute Python code.
The Python interpreter will run within a sandboxed execution environment with some disk space, no Internet access, and limited resources that exist only for the duration of a chat conversation – and subject to a timeout.
Crucially, Code interpreter supports the uploading and downloading of files. Coupled with the fact that subsequent prompts can build on top of previous ones, means it can be utilized to solve a much wider variety of problems than vanilla ChatGPT.
Code Interpreter in action
Users around the world are gushing over it, with at least one data scientist claiming that Code interpreter has saved him hours of work.
Based on various posts on Twitter, some examples of what Code interpreter can do includes:
- Convert image file to MP4: Riley Goodside tweeted about how he uploaded a GIF file and prompted ChatGPT to create a short MP4 video that slowly zoomed in.
- Create an image populated with data from a CSV file: Ethan Mollick loaded a CSV file with the coordinates of lighthouse locations in the US, and got Code interpreter to create an image with them on a map in seconds.
- Analyze your music preferences: Curious about your musical tastes? How about exporting your Spotify playlist and summarizing it using multidimensional PCA and t-SNE analyses? ChatGPT can even produce the code to pull your data from the Spotify API to boot.
- Solving problems using code: Or just get ChatGPT to write the Python code for challenging problems – and run it.
Other abilities mentioned include uploading images containing text for on-the-fly OCR, analyzing text in PDF files, or creating graphs from downloaded datasets of trades or cryptocurrency prices.
The breadth of capabilities has some reports wondering if ChatGPT could replace data scientists. I deem it unlikely, but Code interpreter can certainly save data scientists time by writing code for data handling and quickly completing certain types of activities.
For now, the service is available as an opt-in feature for its Plus users.
Paul Mah is the editor of DSAITrends. A former system administrator, programmer, and IT lecturer, he enjoys writing both code and prose. You can reach him at [email protected].
Image credit: iStockphoto/Svetlana Ivanova
Paul Mah
Paul Mah is the editor of DSAITrends, where he report on the latest developments in data science and AI. A former system administrator, programmer, and IT lecturer, he enjoys writing both code and prose.