Eliminates the resource-heavy teacher+critic distillation stack (Distribution Matching Distillation) used for few-step causal video diffusion, replacing it with generator-only post-training…
Reworks world-action models for robot control so they predict sparse visual deltas plus actions instead of dense future frames, decoupling…
Proposes an RL post-training framework for joint audio-video diffusion generation that fixes credit-assignment problems from entangled multimodal rewards and the…
Instinct, a startup building a personal AI agent that completes real-world tasks such as booking travel, paying bills, canceling subscriptions…
Meta announced the Meta Enterprise Platform, a new business division that packages its AI stack — including the Muse personal…
Elon Musk said xAI’s Colossus 2 data center near Memphis, Tennessee, currently running roughly 110,000 Nvidia GB200 and 440,000 GB300…
Aniket Shah, Jefferies’ head of sustainability and energy-transition strategy, argues that Wall Street’s growing fixation on AI data centers’ carbon…
Med Kharbach and Johanathan Woodworth of Mount Saint Vincent University examine the EU AI Act’s Article 50 requirement, taking full…
Emma Kondrup, Anne Imouza, Catherine Regis and Reihaneh Rabbany analyze Canada’s newly passed Bill C-16 (Protecting Victims Act), which for…
Researchers propose ‘disaggregated quantization,’ which applies distinct quantization schemes to the prefill and decode phases of LLM inference instead of…
The paper introduces PISA, a hierarchical block-sparse attention mechanism that replaces exhaustive query-block scoring with coarse-to-fine selection over a hierarchy…
The paper identifies a credit-assignment flaw in RL training for tool-calling LLM agents: because agent outputs mix structured tool calls…
