Tootfinder

To be clear, re several replies:
(1) The post above is not about •my• opinion on the matter. It’s about which legal arguments make sense.
(2) It’s possible for it to be both: it’s a coherent argument to say that some things an LLM does are in fact literal enough reproduction to be copyright violation •and• to say that humans are responsible for the tool. (That is in fact my own position.) The point of the previous post is that while saying “both!” is logically tenable, saying “neither!” is not.

So... First, this is a great example of the enemy of your enemy not being your friend:
https://todon.eu/@radical_chic/117427249480571937
I can hope that this will cause Adobe to help her the AI bubble popped sooner and/or that it breaks Adobe's market share, but I'm not going to use these tools, nor will I cheer too loudly if the result is that copyright restrictions actually get applied and stop LLM coding (even though stopping LLM coding is unambiguously good). (This is very much not a "clean room" implementation if Claude's training data includes any Adobe source, which it very well might.
Second, assuming this does not cause courts to suddenly start actually applying copyright law correctly, it represents a terrifying new reality for legacy software companies, given their dependence on software monopolies for a lot of their income. The people who have been building on open source all along have the better business model here, but this becoming the norm is more destructive than helpful to the open source movement, since "just rebuild it" seems easier than actually using or contributing to an existing open-source project.
That said, these tools are certainly full of serious bugs. They would be even if they had been written by humans in this timeframe, although in that case you'd probably have an order of magnitude fewer issues. These bugs will result in things like "if you open this PDF with the tool, your computer gets taken over by a virus". That occasionally happens with Adobe products too, of course, but it will happen more frequently and won't be fixed as fast/well in these tools.

To be clear, re several replies:
(1) The post above is not about •my• opinion on the matter. It’s about which legal arguments make sense.
(2) It’s possible for it to be both: it’s a coherent argument to say that some things an LLM does are in fact literal enough reproduction to be copyright violation •and• to say that humans are responsible for the tool. (That is in fact my own position.) The point of the previous post is that while saying “both!” is logically tenable, saying “neither!” is not.