* * * * *
                                        
                      Is this a way of doing AI ethically?
                                        
It was very interesting to see how Corridor Crew [1] used AI (Artificial
Intelligence) to actually solve a work flow issue they had with green screens
[2]—they still had a lot of work to clean up composited footage. They then
worked to train an AI to do that tedious work for them. I don't find an issue
what-so-ever with how they did it—they generated their own training data (so
they avoided the whole scale “stealing” of copyright materal) on their own
equipment (so conceivably they avoided any environmental impact not only by
avoiding a small town's worth of electricity but also by narrowly restricting
the training data to what they absolutely needed) and once it worked, they
are releasing the tool for use by anyone (so avoiding the whole rent-seeking
issues by the major AI players).

I feel like this is how you do AI correctly, by finding a problem that is
tediously repetative and getting the computer to do that work. Now my
question is—what problem is the use of AI-assisted programming actually
solving?

[1] https://www.youtube.com/@CorridorCrew
[2] https://www.youtube.com/watch?v=3Ploi723hg4

Email Sean Conner at sean@conman.org

.