2 ms·

> "engineers optimizing inferencing" They finally implemented DeepSeek open source methods for fast inference?