3 ms·
...and for the people whose hard work this was trained on without permission or compensation?
by wildrhythms 3y ago
...and for the people whose hard work this was trained on without permission or compensation?
- ppsreejith 3y agoAgree that training without permission and compensation is wrong. My argument is more that this tool is a net good. If suppose in the future, Someone releases a similar model trained on purely licensed data (Eg: Like Meta did with their Cm3leon model[0]), my arguments still stand that it gives people more choice and allows them to express more. [0] https://ai.meta.com/blog/generative-ai-text-images-cm3leon/ https://ai.meta.com/blog/generative-ai-text-images-cm3leon/ > As such, and as described in our paper, we’ve trained CM3leon using a licensed dataset.