> Tom Cruise finished filming Top Gun: Maverick years ago, but people are still streaming, renting, and purchasing that movie bringing in massive profit for the studio.

This is also true of the original Top Gun, released ~37 years ago. Very little software has that kind of long-term earning power, further supporting the idea that software work is not all that unique as work goes.

It is as of yet unclear if software will reach that long term profitability. Video games certainly can have it though. I've purchased copies of games which were coded before I was born. Tech is still a young and growing industry where we find new use cases every day. As this slows and as software is perfected this may change. If someone designs the perfect email app it may just be able to run in steady state for decades. We haven't seen that but it may be a reality at some point in the next two decades.

Video games are probably closest to books; in fact, much of software is probably "bookish" in general - lots of it written for particular purposes, sells well enough to have been done, disappears into the long-tail.

A few major breakout successes become historical and bought long after the fact, but the majority do not.

>sells well enough to have been done

Most books almost certainly lose money for the publisher. It's more complicated from the author's perspective given that people write books for a variety of motivations but, certainly, most books are doing well to earn out their advance which can easily be only $1,000 or so.

But, as you say, even those that sell "well enough" initially fall off pretty quickly. And some sorts of titles such as non-fiction about current tech stacks or software versions have a very limited shelf life.

The residual model is very interesting for tech... It's complicated by software refractors over time (kinda leading to compensation questions similar to those around genAI)

I actually never thought of it like this. Thanks for the analog.