genxy
Born on February 06, 2026•364 Karma
>API Services . If you use the API services, we will collect your IP address and the content (text, audio, video, picture) you submit to analyze the relevant instructions based on the model you select and to generate the returned content. Xiaomi will not use the content you provide for model training or any other purposes.
It will come to light that one or many of the Frontier providers held the data, changed ToS and trained later minimally. But I think they just don't care and will train regardless. None of them abide by any level of ethics that would actually prevent them from leveraging an opportunity.
Besides, hasn't SCotUS ruled that raw LLM output isn't subject to copyright? So these companies would be breaking a ToS at worst.
This has been the strategy for months now
A more appropriate title would have been something like "Offline Recurrent Memory Consolidation for Long-Context Language Models". This is supposed to be a research paper, not a story book. The title should give context to other researchers, and not be clearly engineered for clicks. If you don't think so, that's your prerogative, but you're objectively wrong.
It is similar to saying that most websites are just cloud-hosted SQL rebranded.