3 ms·
The moat is a stolen copyrighted data, used for training LLMs. For now you can probably repeat all what OpenLLM did and others too, but the window of opportunit
by Yizahi 2y ago
The moat is a stolen copyrighted data, used for training LLMs. For now you can probably repeat all what OpenLLM did and others too, but the window of opportunity is closing fast. Some companies are making 1-1 deals with existing LLM provides regarding their data, some are closing down, laws are being written slowly but surely.
- 93po 2y agoCan we stop calling it stolen? It wasn't stolen. Copyright infringement is never stealing and it's not even well established infringement is happening.
- Yizahi 2y agoI once thought like that too, "we need balanced, moderate approach etc.". But seeing blatant copyright violations and abuse by megacorps I tend to skew to calling it theft. Purely due to the imbalance of power between individuals and megacorps. The problem we have is that hearing "copyright infringement" people think of different aspects of it. Most people think about how individuals infringe on corpos, like pirating software or a movie and that is indeed is morally grey zone. I was talking purely about reverse situation - when corpos infringe on the copyright of individuals. That is unquestionably abuse of power and in my opinion it is theft. This part at least.