Putting this in front of a kid is twisted.

Adults may justify it for themselves as a comforting fiction, but exposing a child to AI Grandma is waaay over the line.

I think we need a new category and balancing test(s) to handle the breath of generated content.

If I generate a million stories, I should not have full rights over their contents. It's different if I write 99% a story.

What's the line, and does it matter which part is generated? I don't know.

2/2

The existing system wasn't designed for LLMs, but that was just as true for photography and audio recordings. A lot of time and effort has gone into making the existing system robust, and I don't think we'd get a stronger foundation by starting from scratch.

1/2

Also notable, the people who seem to get the most out of it are senior engineers *who get to choose when and how they use it*.

Contrasts sharply with the recent stories of Microsoft employees being forced to use half baked tools or be fired.

I don't think "AI creations cannot be copyrighted" is a study barrier to reverse-centaurization though.

It's a step in the right direction, but the dividing line is going to be created by the same system that made today's IP laws.

Am I going to keep bringing up Cocomelon where it isn't directly relevant? Probably.

If Cocomelon has no haters you'll know I'm truly dead.

Lot of focus on the date here, but a 2 hour party running until 7:30 is also weird at 4. Don't think we've seen any *ending* at 5:30 or later, and the only 2 hour party had them absolutely fried by the end.

People say AI isn't original, but I've never seen a human architect plan for a "Jack & Jil Fun Bah".

I don't know what else there is to say at this point. Openly, gleefully committing war crimes *should* be beyond the pale, yet here we are.