Meta employees have discussed training the company’s AI models using copyrighted works obtained through questionable means, sparking a legal dispute with authors like Sarah Silverman and Ta-Nehisi Coates. Internal chats reveal discussions about buying e-books at retail prices instead of licensing deals with publishers, considering Libgen as a source despite its legal issues, and tuning models to avoid IP risks. Meta’s approach to using copyrighted data for AI training and the potential use of Reddit data are also highlighted in the court documents.
Full Article
Nvidia confirms ‘rare’ RTX 5090 and 5070 Ti manufacturing issue
We have identified a rare issue affecting less than 0.5% of GeForce RTX 5090 / 5090D and 5070 Ti GPUs with one fewer ROP than specified, resulting in a 4% graphical performance impact but no effect on AI or Compute workloads. Affected consumers can contact the board manufacturer for a replacement, as the production anomaly has been corrected. Full Article
Read more