Ben Affleck on Python, tensors and machine learning in visual effects

A director's path into code

Ben Affleck said he has always been kind of into computers since he was young, and that his interest grew when film started to move from analog film to digital. In a quotation collected by Simon Willison and posted on 7th October 2026, Affleck described himself as a Pythonista and said he can write rough Python scripts.

Machine learning in visual effects

Affleck said the visual effects workflow for many years has included machine learning. He placed convolutional neural networks as the sort of precursors to what the transformer can do, which he described as much more computation simultaneously.

He explained that a tensor is the numerical translation of a visual image in numbers, including the batch number, the frame number and the red, green and blue values of each pixel in each frame. A convolutional neural network can identify patterns in that data to reveal edge detection or feature extraction, meaning it can spot enough structure to know where something such as a window ledge is.

Why the patterns matter

The practical point, in his account, is that those identified patterns make it easier to take the green screen image out and replace it with something. The quotation was tagged machine learning and Python on Willison's weblog.

Sources: Simonwillison