[veunes]

Nah, those are completely different beasts. DeepSeek's MLA solves the KV cache issue via low-rank projection - they literally squeeze the matrix through a latent vector at train time. TurboQuant is just Post-Training Quantization where they mathematically compress existing weights and activations using polar coordinates

[esafak]

No, it is about compressing the KV cache; see How TurboQuant works.