I don't have, or want to have deep extensive thoughts.
The major commercial legitimate monetize-able use cases will be around the margins at the start, until the models are good enough to be reliable. Because really you can't automate away with something you cannot trust to completely eff up your relationships and stack.
As of right now, the primary use case is to help students cheat with essays, undermine democracy and destroy social media, and undermine artists by copying their styles and selling to their target audience.
There's lots of ethical questions around the provenance of training data, the nature of consent around it, and how this has poisoned future data forever. My prediction is that at consumer-facing level, this will be a lot less of a revolution than the Ai companies are making out to be.
No comments:
Post a Comment
Tell me what you think. I'll read, promise.