zlacker

>>davidg+(OP)
I’ve noticed that people tend to disapprove of AI trained on their profession’s data, but are usually indifferent or positive about other applications of AI.

For example, I know artists who are vehemently against DALL-E, Stable Diffusion, etc. and regard it as stealing, but they view Copilot and GPT-3 as merely useful tools. I also know software devs who are extremely excited about AI art and GPT-3 but are outraged by Copilot.

For myself, I am skeptical of intellectual property in the first place. I say go for it.

>>kweing+v6
I think the distinction is that only one of those classes tends to produce exact copies of work. Programmers get very upset at DALL-E and Stable Diffusion producing exact (and near-exact) copies of artwork too. In contrast to exact copying, production of imitations (not exact copies, but "X in the style of Y") is something that artists have been doing for centuries, and is widely thought of as part of arts education.

For some reason, code seems to lend itself to exact copying by AIs (and also some humans) rather than comprehension and imitation.

>>pclmul+3b
I'm mildly suspicious that this example is an implementation of a generic matrix functionality though: you couldn't patent this sort of work, because it's not patentable - it's a mathematics. It's fundamentally a basic operation, that would have to be implemented with a similar structure regardless of how you do it.

>>XorNot+Ai
Mathematics is not patentable, but you can patent the steps a computer takes to compute the results of that particular algorithm.

>>heavys+ym
Only if it has physical consequences. There was a case in 2014 that narrowed software patents significantly, called "Alice vs CLS Bank." No more patents on computerized shopping carts, but encryption or compression can still be patented.