What if every decode step gave the next one a head start? Meet Guess-Verify-Refine — a new hardware-aware sparse-attention algorithm from NVIDIA Research. Built for TensorRT LLM on Blackwell, it reuses temporal patterns across decode steps for: → 1.88x faster Top-K attention
AI HARDWARE
-
xAI and Anthropic have opposite compute problems despite large GPU fleet
By
–
The xAI / Anthropic compute story is not about one company having GPUs and the other wanting them.
It's that they have opposite problems. xAI reportedly runs one of the largest GPU fleets in the world. Yet according to The Information, its recent model FLOPs utilization was -
Genesis AI se lance dans la robotique full-stack
By
–
Genesis AI pivoted hard this week from model lab to a full-stack robotics company.
— The Rundown AI (@TheRundownAI) 7 mai 2026
The company released its first foundation model, GENE-26.5, alongside a dexterous robotic hand and a data-capture glove.
GENE-26.5 can run a range of robots, including those by other… pic.twitter.com/MbYTTnMoNoGenesis AI pivoted hard this week from model lab to a full-stack robotics company. The company released its first foundation model, GENE-26.5, alongside a dexterous robotic hand and a data-capture glove. GENE-26.5 can run a range of robots, including those by other
-
AI exploits real-world signals
By
–
The third piece is what makes it actually useful. Real-world signals. → AI pulls insights from what people are sharing
→ Adapts answers based on live context
→ Even builds tools on the fly, dashboards, mini apps, simulations And with AI glasses coming… It won’t just -
QUEBEC.AI launches AGI Alpha Nodes and AI-first Infrastructure
By
–
http://
QUEBEC.AI launches Agents & Infrastructure: AGI Alpha Nodes for governed machine work at massive scale. http://
QUEBEC.AI lance l’infrastructure AI‑First souveraine: nœuds AGI Alpha, preuve, validation, gouvernance. https://
quebecartificialintelligence.com/agents-infrast
ructure
… #QuebecAI -
Zyphra’s ZAYA1-8B delivers high reasoning with under 1B parameters
By
–
Zyphra under 1B active parameters, AMD-Trained, big evals, look strong? Zyphra says its new ZAYA1-8B model delivers unusually high reasoning power for its size, using under 1 billion (!) active parameters while competing with much larger open-weight and proprietary systems on
-
Top AI stories: partnerships, trials and tools
By
–
Top stories in AI today: – Anthropic, SpaceX partner in new compute deal
– Mira Murati speaks out in Musk vs. OpenAI trial
– Use Claude Design’s slide decks feature like a pro
– DeepMind picks EVE Online game as next AI testbed
– 4 new AI tools, community workflows, and more -
Elon Musk announces multiple Grok models training in Colossus 2
By
–
We have many great Grok models training simultaneously in Colossus 2. I am confident that they will bear fruit. And work on the Grok Built harness is progressing well.
-
Anthropic’s Potential with NVIDIA GPUs and Power
By
–
Okay Anthropic, show us what you could do with 220,000 NVIDIA GPUs and 310MW.
