@adlrocha - What if AI doesn’t need more RAM but better math?
adlrocha.substack.com
How TurboQuant compresses the KV cache without losing accuracy, and what that could mean for memory stocks