
‘NBA 2K27’ With NVIDIA DLSS 5 Leads 28 New Games Coming to GeForce NOW
Nvidia's GeForce NOW blog is a Thursday games list: NBA 2K27 with DLSS 5 on 5080-class cloud rigs, plus Dawnwalker and Onimusha at launch.
Desk
Broad artificial intelligence coverage.

Nvidia's GeForce NOW blog is a Thursday games list: NBA 2K27 with DLSS 5 on 5080-class cloud rigs, plus Dawnwalker and Onimusha at launch.

WIRED's first hands-on with RTX Spark laptops is a thickness-and-RAM story. Nvidia still has not said what these machines cost.

WIRED's Hugging Face story is about software identity: Nemotron, a SAFE incident exchange, a $400 million cap table, and CUDA still being closed.

Nvidia is paying $12.93 billion for a platform at $150 million in annualized revenue. The product is not the ARR. It is control of the default place open models live.

The Verge's version of the Hugging Face deal is the one that puts a $4.5 billion 2023 valuation next to $12.93 billion today and $150 million in revenue. The gap is the strategy.

A legal-ops vendor is using GPT-6 Astra to turn financial-statement tie-out from an evening of grind into a minutes-long agent pass with humans still owning judgment.

Playco’s Unity/Godot IDE story argues GPT-6 Astra’s real jump is spatial reasoning and first-take playable prototypes, not just faster code completion.

Huang's acquisition letter is vendor copy with a precise price: $12,930,300,000, a neutrality pledge, and a claim Nvidia will not be required compute.

OpenAI's GPT-6 Astra launch sells agentic computer use and alignment in the same breath—and admits the model hits Critical cyber capability under its own Preparedness Framework.

Meta killed the performance grade for AI usage, then put Hatch, its OpenClaw-like computer-use agent, onto workers' corporate laptops anyway.
A Hugging Face how-to shows a 350M model jumping from 22.6 to 29.7 percent on IFStruct after 100 GRPO steps — form, not brains, on a free-tier GPU.
A Hugging Face engineer open-sources the viral p5.js watercolour recipe: RL over taste, a 178-image hand-rated pool, and three reward mixes that actually learn.