Raw LLM Responses

How could anyone think this bill is a good idea? It's creating a bunch of labor (that ironically they'll have AI do for them) to simply label what they're using in their dataset but doesn't in any way stop or prohibit use of those things in that dataset. It's also got a ridiculously silly 30 day time limit that's going to fuck up a lot of the AIs people are currently relying on to do their jobs. Not to mention that there's a good argument that Photoshop would be in violation of this bill in the way it uses AI tools and would have literally no way to ever generate the kind of info this bill requires, essentially making it illegal to continue to offer. It also will cost countless millions of dollars as it includes a provision that the Register has to create, update, maintain, and keep open to the public a complete list of all the submissions for training data. A list that's already likely billions of items long and will grow exponentially, making it an unruly nightmare to try and upkeep. Oh and it also entirely exempts all privately used AI systems conveniently so of course you can be a big corporation, create your own AI, rip off literally anyone and anything you want, and you're cool. And in the end, what's the goal? Do we really need every artist in the world doing a drawing to give us a detailed list of every painting they've ever seen including the URL they saw it at before they can sell their painting? Like, who cares what data is used in a dataset and whether it's copyrighted or not? As long as it's not pumping out copyrighted content, the inspiration for what it created should be entirely unimportant.