Pinned post

As inference splits into prefill and decode, Nvidia's Groq deal could enable a "Rubin SRAM" variant optimized for ultra-low latency agentic reasoning workloads (Gavin Baker/@gavinsbaker)

Gavin Baker / @gavinsbaker : As inference splits into prefill and decode, Nvidia's Groq deal could enable a “Rubin SRAM” variant optim...

17 January 2025

EU expands its X probe, asking for internal docs about X's recommendation algorithm, following complaints from German politicians about boosting the far right (Financial Times)

Financial Times:
EU expands its X probe, asking for internal docs about X's recommendation algorithm, following complaints from German politicians about boosting the far right  —  Move follows complaints from German politicians ahead of election next month  —  Brussels has ordered Elon Musk …

Posted from: this blog via Microsoft Power Automate.

Daily Deals