Movies are an easy example: Do you want to watch the original ending of Return of the Jedi, the special edition ending, the revised ending post-prequels, the 4:3 edition or the widescreen edition? There are plenty of reasons to want to keep a copy of the version of the experience you enjoyed and want to experience again.

In that example, I'd wager most people wouldn't care which version it is.

Movies still come out on disc because they're good stocking stuffers for the holidays, and you can carve out a pretty good margin selling box sets of film anthologies.

Plus Blu-rays can be sold globally. What's available in one country's Netflix often doesn't line up with what's available in a different country. Plus, a good chunk of the US movie-watching population does not have reliable broadband access.

Game developers would have more to gain from a streaming-only model:

1. It's expensive to develop for multiple platforms. If you can get people to buy into streaming, you theoretically have only one hardware target.

2. Streaming is more effective at enforcing DRM for games. You can make a recording of a movie you're playing on Netflix, but games would require remote execution which can't be copied.

>but games would require remote execution which can't be copied.

Next HN headline: Machine Learning algorithm reverse engineers a Stadia game.

I think there has been some reverse engineering of simple game engines through ML already actually.